Communication facilitation methods, programs, computers, AI-enabled chat service providers, and information processing devices.

Abstract

AI is evolving to learn various personal information and better represent a person's personality and characteristics. Attempts are also being made to utilize such AI to improve human-to-human communication. However, there remains a challenge in effectively using AI to facilitate smoother communication and better human interaction. A personalized AI is used to speak on behalf of the individual. By having the AI ​​speak, it creates a flow of conversation. The individual can observe and listen to the conversation, and at some point, they can take over from the AI ​​to continue the conversation, or join the conversation as a new guest. This makes it easier to construct a conversation than building one from scratch, facilitating smoother communication and interaction between people.

Description

Technical Field

[0001] Technologies that incorporate an individual's personality, characteristics, preferences, etc. into AI for personalization are being studied in various places, and various methods have been developed. How to utilize such personalized AI has also been under consideration.

Background Art

[0002] Personalization that incorporates an individual's personality, characteristics, preferences, etc. into AI to create another self has been studied. There are numerous studies that attempt to create a perfect individual by training an AI model with an individual's personality and characteristics. Even if it is not aimed at such perfection, there are also simpler ways to personalize AI. We have entered an era where AI can be easily used by individuals, and there are web pages where you can easily chat with AI. When such a page is requested to teach the AI an individual's personality, characteristics, etc. and receive responses accordingly, it is possible to receive responses in a form that follows the request to a certain extent. This can also be regarded as one form of personalization. Research is also underway to utilize these personalized AIs to assist people in communication.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] As described in Patent Document 1, AI is evolving to learn various personal information and better represent a person's personality and characteristics. Attempts are also being made to utilize such AI, which mimics an individual's personality and characteristics, to improve person-to-person communication. However, there is a challenge in that AI has not yet been effectively used to facilitate smooth communication or person-to-person interaction. [Means for solving the problem]

[0005] This system uses methods such as having an AI machine-learn to create a model of an individual's personality and characteristics, providing the AI ​​with information about the individual's personality and characteristics to reproduce them, or combining these two methods to reproduce the individual's personality and characteristics, allowing the "personalized AI" to speak on behalf of that individual. The AI's speech flows, creating a conversational flow. The individual can observe and listen to this conversation, taking over from the AI ​​midway through, or joining the conversation as a guest, making it easier to construct a conversation than starting from scratch. Furthermore, the individual can converse with other "personalized AIs." Since the other party is an AI that has registered the individual's personality or has been trained to reflect the individual's personality, the conversation is more relaxed and less formal than talking to a real person. Conversations can even take place in online chat spaces without meeting in person. Although you are talking to a "personalized AI" and not the person themselves, the conversation with such an AI reflects the original person's personality, characteristics, habits, and preferences. Therefore, you can maintain connections by registering the IDs of AIs you enjoy talking to or AIs that seem to share your interests. In this way, it is possible to facilitate communication between people and efficiently expand interactions. [Brief explanation of the drawing]

[0006] [Figure 1] Image of AI management function [Modes for carrying out the invention]

[0007] Personalization through AI is becoming a common technology. For example, there are already services that allow people connected to a deceased person to communicate with the AI ​​after their death, making them feel as if they are talking to the deceased. In this way, the technology of transferring an individual's personality and characteristics into an AI and making it behave like that person, even if not perfectly, is becoming commonplace.

[0008] The personality and characteristics of an AI, which are created by training or providing information to the AI, may belong to the administrator managing it, to someone else, or to a fictional character. However, if the administrator uses the personality or characteristics of someone else, it is preferable to obtain appropriate permission from the person or rights holder of that personality or characteristics.

[0009] In the present invention, AI (artificial intelligence) refers to a computer system or algorithm that imitates, extends, or complements some or all of the intellectual capabilities of humans. This includes, but is not limited to, machine learning, deep learning, reinforcement learning, neural networks, expert systems, natural language processing, computer vision, speech recognition, speech synthesis, and decision support systems. In the present invention, "personalized AI" refers to a state in which an individual's model and information are applied to a language model AI such as LLM or a speech recognition / speech synthesis AI to reproduce the original individual's personality, characteristics, voice, etc., and to enable conversation. However, in the present invention, it is not necessary to use a language model AI such as LLM or a speech recognition / speech synthesis AI as long as natural conversation can be achieved; any AI may be used.

[0010] In the present invention, conversation and chat refer to conversations conducted via voice or text, or a combination of both, unless otherwise specified.

[0011] The server described in this invention may be any server unless otherwise specified. It may be a cloud server or an on-premise server. Any server on which this invention can be executed is acceptable.

[0012] When using AI for conversation, a large-scale language model (LLM) is suitable. LLMs learn from large amounts of text data, understand context, and can generate natural and meaningful responses. However, in the present invention, a large-scale language model (LLM) is not required, and any AI can be used as long as a natural conversation can be achieved.

[0013] Such AI can be trained to learn an individual's characteristics and personality, and to some extent, incorporate that individual's way of thinking and statements. Furthermore, the results can be saved as models or information. This can be saved as models, or in other forms such as parameters, rules, and settings. Both may be used. By saving these models and information and applying them to the AI ​​as needed, it becomes possible to have conversations that are always tailored to the individual's characteristics and personality.

[0014] Furthermore, these AIs can mimic an individual's characteristics and personality to a certain extent simply by providing them with information about that individual's traits and personality. For example, there are web-based AI-powered websites like Google Gemini and Microsoft Copilot where you can converse with an AI. On these websites, if you command the AI ​​to "converse as if you were a celebrity (any well-known celebrity)," it can easily mimic that celebrity. This is done by the AI ​​inferring the words and manner of speaking from the celebrity's information it already possesses. The AI ​​can also converse according to commands such as "converse as if you were a nervous person" or "converse as if you were an optimistic 30-year-old man." From this, it can be seen that simply providing an AI with information about an individual's characteristics and personality allows it to converse to a certain extent in line with that person's traits and personality. In this way, by applying models and information to an AI, it is possible to reproduce that individual's characteristics and personality.

[0015] Furthermore, if you want to conduct these conversations using voice, you can use speech synthesis software in combination with other methods for uttering words. For listening to speech, you can use speech recognition software in combination with these methods. While LLMs can mimic individual conversational features (vocabulary, style, phrase usage, etc.) at the text level, and can learn the speaking and writing styles of specific individuals to generate similar texts, they cannot perform speech recognition and synthesis. LLMs typically receive text input and generate text output, but speech recognition and speech synthesis need to be handled by separate systems or models. By utilizing speech-to-text and text-to-speech synthesis systems, voice conversation becomes possible.

[0016] In the field of speech-to-text systems, some are using AI to improve recognition technology. Whisper (a speech recognition AI) is one such example, and it is also possible to combine speech recognition systems that utilize these AIs. In the present invention, any AI can be used as long as natural conversation can be achieved.

[0017] In text-to-speech (TST), if you want to express the characteristics of an individual's voice quality and speaking style, you can use "speech synthesis AI" or "speech cloning AI" to generate speech. Similar to the previous example, by applying models and information, you can create conversations that reflect an individual's characteristics and personality. An example of this is WaveNet (a deep learning model for generating speech waveforms). In the present invention, any AI can be used as long as it enables natural conversation.

[0018] For example, combining these AIs—Whisper (speech recognition AI), ChatGPT (language modeling AI), and WaveNet (speech synthesis AI)—can be used for voice conversations.

[0019] Thus, it is a common technique to combine and use language model AIs such as "voice synthesis AI", "voice cloning AI", and LLM with voice recognition AIs. In the present invention, these "voice recognition AI", "voice synthesis AI", and "voice cloning AI" are expressed as voice recognition / voice synthesis AI.

[0020] When using these various AIs, it is possible to borrow external AIs via APIs using cloud services, etc., or it is also possible to set up and use dedicated servers or computers by oneself.

[0021] The technology of personalizing AI is still under research, and its development is expected. The "personalized AI" in the present invention is one that reproduces an individual's personality and characteristics, and any AI or mechanism can be used as long as it has the mechanism. The "AI management function" manages it.

[0022] The "AI management function" is software that manages personal models and information applicable to language model AIs such as these LLMs and voice recognition / voice synthesis AIs, connects to language model AIs such as LLMs and voice recognition / voice synthesis AIs that use personal models and information, and enables conversation. Multiple personal models and information can be managed, and it also has a function to select the ones the administrator wants to use each time.

[0023] The "AI management function" has a function to manage personal models and information. The "AI management function" pre-stores information on how to apply which models and information to language model AIs such as LLMs and voice recognition / voice synthesis AIs to be used.

[0024] That is, the "AI management function" automatically connects via an API or the like using the connection information to the AI according to the requirements of the administrator, and applies it to language model AIs such as LLMs that use the models and information managed by the "AI management function", or voice recognition / synthesis AIs, so as to enable conversations that reproduce an individual's personality, characteristics, voice, etc. The "AI management function" has a function that can be applied to language model AIs such as LLMs or voice recognition / synthesis AIs that use an individual's models and information to enable conversations. And the "AI management function" also stores the connection information to language model AIs such as LLMs or voice recognition / synthesis AIs that are necessary to enable conversations according to the requirements of the administrator.

[0025] Such mechanisms already exist. In Microsoft Azure Machine Learning + Azure Bot Service, a model trained in Azure Machine Learning can be deployed and used in Azure Bot Service to build chatbots and conversational AIs. In AWS SageMaker + Amazon Lex, a model is hosted in SageMaker and its endpoint is incorporated into Lex to realize natural conversations. Lex supports voice and text-based interfaces.

[0026] Figure 1 is a diagram showing the application to language model AIs such as LLMs or voice recognition / synthesis AIs that use the models and information of individuals managed by the "AI management function (1)".

[0027] The "AI management function" can operate as a server, connected to and operated by a computer such as a mobile phone, or it can run directly on the computer such as a mobile phone for conversation. It can also run on the same server or computer as the language model AI (LLM, etc.) or speech recognition / speech synthesis AI being used. Whether the "AI management function" is running as a server or on a computer such as a mobile phone, the administrator can perform various operations through a dedicated web page or application. Furthermore, when conversing with the "personalized AI" using text instead of voice, the conversation of the "personalized AI" can be displayed on a dedicated web page or application, and replies can also be entered as text from that web page or application.

[0028] Administrators can initiate a conversation with a personalized AI by connecting to the server from a computer, such as a mobile phone, via these dedicated web pages or applications, or by activating the "AI management function" if it is already installed on the computer, such as a mobile phone.

[0029] The "AI management function" allows for the application of individual models and information to the AI, enabling a "personalized AI" to converse. This allows administrators and other humans to interact with the "personalized AI." Similarly, it's possible to have a "personalized AI" converse with someone else's managed "personalized AI." Administrators and other humans can speak to the "personalized AI" via voice or type in a text message to provide a topic (such as today's news), which can then be used to provide conversation topics. This allows the "personalized AI" to assist in conversations when human conversations stall.

[0030] Because "personalized AI" speaks in accordance with the personality and characteristics of the individual it is based on, administrators can predict the content and flow of the conversation to some extent, allowing them to listen and observe with peace of mind. Furthermore, it's possible to have an administrator's "personalized AI" converse with another person's "personalized AI," and then, when the conversation becomes interesting, the administrator and others can join in.

[0031] When a "personalized AI" communicates with an administrator or other human, voice communication allows for the transmission and reception of conversation audio through the speaker and microphone of the computer running the dedicated "AI management function" web page or application. Text communication allows for both text display and text input on the dedicated "AI management function" web page or application. Administrators and other humans can then input text into this dedicated web page or application, and the "personalized AI's" response will be displayed and viewable as text.

[0032] When an administrator's "personalized AI" and another "personalized AI" from a different "AI management function" communicate via voice, they must be able to transmit and receive audio from their respective speakers and microphones. This allows for mutual input and output of the conversation, enabling conversation. For text-based conversations, the "AI management function" has a dedicated web page or application that allows for both text display and input of the conversation. Therefore, another person must act as an intermediary, inputting the text into the computer running the respective "AI management function" web page or application. When one "personalized AI" displays the conversation, the content is input as text into the computer running the other "personalized AI"'s "AI management function" web page or application. This process is repeated.

[0033] The "AI management function" has the ability to connect with and communicate with other "AI management functions" using communication methods such as internet communication, wireless, and wired communication.

[0034] This communication can either involve checking the connection settings of each other's "AI management functions" and using a mechanism to directly connect the "AI management functions" to each other in a peer-to-peer manner, or it can be done by preparing a dedicated server for the "AI management functions" and exchanging IDs through that server, similar to messengers like WhatsApp.

[0035] This allows administrators to communicate using the "personalized AI" of the "AI management function" as if they were on a phone call, even when they are geographically separated. Furthermore, even when administrators are in close proximity, text conversations are conveniently linked to each other's "AI management functions," eliminating the need for a human to mediate the text conversation as mentioned earlier.

[0036] The "AI management function" allows administrators to receive requests and commands for the "personalized AI" via voice or text through a dedicated web page or application. The "AI management function" communicates these requests and commands to the language model AI (such as LLM) or speech recognition / speech synthesis AI being used. This is done by adding the requests and commands to the conversation.

[0037] Administrators can listen to the conversation of the "personalized AI" and, through a dedicated web page or application for "AI management functions," issue requests and commands, guiding the conversation in the direction the administrator prefers. Since the administrator is not participating in the conversation themselves, they can manipulate the conversation as if they were a third party listening. When the administrator issues such requests or commands, the "AI management function" transmits those requests and commands not only to the conversation but also to the language model AI (such as LLM) and speech recognition / speech synthesis AI being used.

[0038] This allows administrators to make requests and commands to the "AI management function" via voice or text while watching or listening to the "personalized AI" conversation, such as "speak faster," "choose your words carefully," "speak a little slower," "ask a riddle," or "end the conversation now." In this way, administrators can control the conversation to their liking without having to speak themselves.

[0039] If the "personalized AI" is communicating via voice, the administrator can input requests and commands as text through a dedicated web page or application for "AI management functions" without interfering with the voice conversation. Conversely, if the "personalized AI" is communicating via text, the administrator can input requests and commands as voice through a dedicated web page or application for "AI management functions" without interfering with the text conversation. This function prevents unnecessary words from being inserted into the conversation. In other words, requests and commands are conveyed to the language model AI (such as LLM) or speech recognition / speech synthesis AI being used without the conversation partner knowing. The dedicated web page or application for "AI management functions" can transmit the administrator's requests and commands to the language model AI (such as LLM) or speech recognition / speech synthesis AI in addition to the current conversation, even if the "personalized AI" is communicating via text. Conversely, even if the "personalized AI" is communicating via voice, it can accept text input and transmit the administrator's requests and commands to the language model AI (such as LLM) or speech recognition / speech synthesis AI in addition to the current conversation.

[0040] If the "personalized AI" communicates via voice, administrators can send requests and commands via text input through a dedicated web page or application for "AI management functions" without interfering with the voice conversation. Conversely, if the "personalized AI" communicates via text, administrators can send requests and commands via voice input through a dedicated web page or application for "AI management functions" without interfering with the text conversation. This means that there is a function to prevent unnecessary words from being introduced into the conversation.

[0041] The "AI Management Function" has the ability to pause and resume conversations of "Personalized AIs." The "AI Management Function" receives requests and commands from the administrator regarding conversations and has the function to transmit those commands to the language model AI (such as LLM) or speech recognition / speech synthesis AI that is using those requests, thus enabling pausing and resumption. In the case of pausing, the "AI Management Function" instructs the language model AI (such as LLM) or speech recognition / speech synthesis AI used by the "Personalized AI" in the conversation to pause and to monitor the conversation during the pause. This instruction can be given by the administrator via text or voice, or a pause button can be provided on a dedicated web page or application for the "AI Management Function," and when pressed, the "AI Management Function" is programmed to instruct the language model AI (such as LLM) or speech recognition / speech synthesis AI to pause and monitor the conversation during the pause. In the case of resuming, the "AI Management Function" instructs the language model AI (such as LLM) or speech recognition / speech synthesis AI used by the "Personalized AI" in the conversation to take over the conversation that was being monitored up to that point and start the conversation on behalf of the administrator. This instruction can be given by the administrator via text or voice, or it can be programmed to include a resume button on a dedicated web page or application for the "AI management function," which, when pressed, instructs the "AI management function" to take over the conversation it was monitoring and start the conversation on behalf of the administrator, such as an LLM or speech recognition / speech synthesis AI.

[0042] This allows the administrator to temporarily pause the conversation of their "personalized AI," and then take over the conversation from where the "personalized AI" left off. If the administrator then wants to let the "personalized AI" continue the conversation, they can simply resume it.

[0043] Thus, the "AI Management Function" can be paused and resumed by providing pause and resume buttons in the "AI Management Function" app or a dedicated web page, or by using voice commands or text input instructions. Furthermore, when paused, the "AI Management Function" will stop displaying the conversation from the "Personalized AI" to the other party if the conversation was previously conducted in text, and will stop the voice output from the "Personalized AI" so that the other party cannot hear the conversation if it was conducted in voice. When resumed, these displays and voice outputs will be restarted.

[0044] If the administrator later loses interest in the conversation, they can resume the "personalized AI" by unpausing it. This allows the "personalized AI" to continue the conversation from where the administrator left off. Since the "personalized AI" monitors the conversation even while it's paused, a smooth handover of the conversation is possible.

[0045] Using pauses in this way makes it easier for administrators to grasp the flow of the conversation and reduces barriers to their participation. Furthermore, even if the conversation veers in a direction the administrator did not want, they can have the "personalized AI" resume the conversation, allowing them to restructure the conversation or have the "personalized AI" close the conversation, which is very convenient.

[0046] While the "personalized AI" is paused or while the administrator is conversing with the "personalized AI," it will observe and listen to the administrator's conversation. The "personalized AI" has the ability to convey its perspective as advice to the administrator through the "AI management function." This is because the "AI management function" receives requests and commands from the administrator regarding the conversation and transmits these requests and commands, in addition to the current conversation, to the language model AI (such as LLM) or speech recognition / speech synthesis AI being used. The "AI management function" can receive advice by telling the language model AI (such as LLM) or speech recognition / speech synthesis AI being used, "Please give advice while the administrator is conversing." This advice can be displayed as text on the "AI management function" app or dedicated web page screen, or conveyed via voice.

[0047] If the "personalized AI" is communicating via voice, the administrator can display advice as text through a dedicated web page or application for "AI management functions" without interfering with the voice conversation. Conversely, if the "personalized AI" is communicating via text, the administrator can output advice as voice through a dedicated web page or application for "AI management functions" without interfering with the text conversation. This function prevents unnecessary words from being inserted into the conversation. In other words, advice can be received without the conversation partner knowing. The dedicated web page or application for "AI management functions" has the function to transmit advice from the language model AI such as LLM and the speech recognition / speech synthesis AI being used, even if the "personalized AI" is communicating via text, through the speaker's voice output, and conversely, to output and transmit advice from the language model AI such as LLM and the speech recognition / speech synthesis AI being used, even if the "personalized AI" is communicating via voice.

[0048] The desired advice for the "AI Management Function" can be saved as a pre-configured setting and incorporated into the "AI Management Function." An advice button can be placed on a dedicated web page or application for the "AI Management Function," similar to the pause function. When this button is pressed, the "AI Management Function" can communicate the need for that advice to language model AI such as LLM or speech recognition / speech synthesis AI. For example, pressing the advice button might convey a message such as, "While the administrator is having this conversation, please text how you would respond." This message can be added to the current conversation. The administrator can also give instructions via voice or text at any time. For example, if the administrator says, "While the administrator is having this conversation, please text how you would respond," the "AI Management Function" will respond with, "I would say that this response is OK," which can then be displayed as text or voiced on the "AI Management Function" app or dedicated web page screen. If the administrator says, "Predict and speak for the other person's feelings," the "AI Management Function" will respond with, "It seems the other person is not comfortable talking about earthquakes," which can then be displayed as text or voiced on the "AI Management Function" app or dedicated web page screen.

[0049] The "AI management function" can save such conversation history to a server or computer such as a mobile phone that manages the "AI management function". It records not only the statements of the administrator's "personalized AI" but also the statements of the person it is talking to. In the case of text conversations, the history is saved as text data. In the case of voice conversations, the history is saved as voice data, or the administrator's "personalized AI" statements are saved as text, and the other person's voice is recognized by speech recognition and recorded as text data in the history, or both types of history can be saved.

[0050] In conversations using these "personalized AIs," if an administrator joins the conversation or pauses the "personalized AI" and takes over the conversation themselves, the "personalized AI" can record the administrator's statements, which can then be used as training material for the models and information used by that "personalized AI."

[0051] The "AI management function" has the ability to change the language used as the basis for conversations. Since the "personalized AI" is a language model AI such as LLM or a speech recognition / speech synthesis AI, it can converse in English or other languages ​​by requesting a change in the language used as the basis for conversations from the "AI management function". The administrator can also learn the language by listening to the conversation or taking turns speaking during it.

[0052] In this way, administrators can access an "AI management function" app or a dedicated web page from their mobile phone or computer and operate it to make their "personalized AI" engage in conversation. Furthermore, the administrator can control the content of the conversation from the "AI management function." This makes it possible to initiate conversations with others and liven up conversations that have stalled. Because administrators can control conversations with few commands, they can manipulate the conversation as if they were operating a machine, creating a new kind of enjoyment. Since administrators are not speaking themselves but are manipulating the conversation, they can observe and listen to the conversation from a third-party perspective, which is expected to reduce their aversion to conversation.

[0053] Furthermore, by using a "personalized AI" as a conversation partner, users can try conversations they wouldn't normally have and discover new sides of the other person. For example, even when debating a controversial topic or discussing something that some people are very passionate about, such as sports or politics, the administrator doesn't participate in the conversation but only manipulates it, allowing them to observe and listen to the conversation from a third-party perspective, thus preventing them from becoming overly emotional and enabling a calmer discussion.

[0054] The "AI management function" has the capability to automatically receive conversation requests for the "personalized AI" from other servers, computers, or services and relay those conversations to those servers, computers, or services.

[0055] Other servers, computers, and services include chat services and messengers such as WhatsApp, personalized AI-enabled chat services, virtual space software and computer games, and programs deployed on social networking services and websites.

[0056] The "AI Management Function" automatically receives conversation requests from other servers, computers, or services for the "Personalized AI" and coordinates those conversations. The administrator can then monitor these conversations using the "AI Management Function." This monitoring function can be accessed through the "AI Management Function" app or a dedicated webpage. If the "AI Management Function" is conducting multiple conversations, the administrator can switch between them to view and listen to them.

[0057] The functions of the "AI Management Function" described above can still be used for conversations being monitored by the "AI Management Function" monitoring function. The monitoring function can also be used for features such as the administrator making requests and giving commands to the "personalized AI" in conversations, pausing and resuming conversations of the "personalized AI," and the ability for the "personalized AI" to communicate its perspective as advice to the administrator through the "AI Management Function."

[0058] The "AI management function" can have features similar to messengers like WhatsApp. This means you can register other "AI management functions" or human accounts and communicate with them whenever you want. Conversations can be done via text or voice.

[0059] To allow individuals without "AI management capabilities" to communicate with other administrators' "AI management capabilities," a dedicated server, app, or webpage capable of communicating with "AI management capabilities" must be prepared. Individuals without "AI management capabilities" can register their account, handle name, or ID through this platform, and then register other "AI management capabilities" or individuals to their account in a contact list to begin conversations.

[0060] The "AI Management Function" allows you to entrust conversations with other registered "Personalized AIs" or humans to the "Personalized AI." It has a function that allows the administrator to set whether to handle conversations themselves or have the "Personalized AI" handle them when a conversation request is made or when a conversation is requested. This setting can be switched at any time. When the "Personalized AI" is set to handle conversations, the "Personalized AI" managed by the "AI Management Function" will automatically engage in conversations when a conversation request is made from another "Personalized AI" or human account, or when you yourself request a conversation.

[0061] When conversations are entrusted to a "personalized AI" in this way, the "AI management function" allows you to pre-specify conversation settings such as speaking style, conversation content, and how the conversation should proceed. For example, "Speak calmly," "Talk about today's baseball game," or "Start with a soft conversation and become more aggressive in the middle of the conversation."

[0062] It also has a function that allows you to specify the speaking style, conversation content, and how to converse for each specific handle name or ID. Furthermore, it has a function to pre-register messages for specific handle names or IDs, and when a conversation request is received from that handle name or ID, the message can be sent.

[0063] It also has a calendar function for registering appointments. For example, you can set it up so that from 9 PM to 8 AM, if a conversation request comes in from another "AI management function" or a human, the "personalized AI" will respond, and at other times, the administrator will be notified. Furthermore, you can specify the speaking style, content, and manner of conversation for each contacting username or ID, such as "speak concisely" or "speak politely," and you can also leave specific messages for each contacting username or ID. You can also set the conversation method and messages to be changed at specific times or dates.

[0064] Of course, administrators can always respond to requests for conversation themselves.

[0065] Additionally, this calendar has a function that allows administrators to schedule other appointments unrelated to the conversation.

[0066] The "AI management function" may have the group chat functionality found in other messengers. The administrator can participate in the group chat themselves, but can also schedule messages from the "AI management function" in the calendar. For example, from 9 PM to 8 AM, if another "AI management function" or a human contacts the group chat, it may send a message such as "I will respond tomorrow morning." Or, if a specific username or ID contacts the group chat on a specific day, it may send a specific message to the group chat.

[0067] Whether a conversation request is made or initiated by the user, the functions of the "AI management function" described above remain available, regardless of whether the administrator or the "personalized AI" responds. The administrator can make requests and commands regarding the "personalized AI's" conversation, pause and resume the conversation, and communicate the "personalized AI's" perspective as advice to the administrator through the "AI management function." These functions can still be used as a messenger.

[0068] It's possible to use existing chat services designed for general person-to-person conversations, or to have a "personalized AI" converse through existing messengers such as WhatsApp. In this case, you log in to the chat service or messenger like WhatsApp using an account provided by the administrator, and proceed with the operation until you can converse with the other party. If it's a voice conversation, the other party's speech is responded to through the computer's speakers, so you input that into the microphone of the computer running the "AI Management Function" app or dedicated web page, and the "AI Management Function" then transmits the speech to the language model AI (such as LLM) or speech recognition / speech synthesis AI being used. The "AI Management Function" app or dedicated web page then responds with voice from the computer's speakers, which you input directly into the microphone of the computer connected to the chat service or messenger like WhatsApp. By repeating this process, a conversation can be established.

[0069] Alternatively, if a chat service or messenger like WhatsApp has an API that allows it to share conversation content with other software, the "AI management function" can use it to acquire the voice conversation, then pass the voice conversation through the "AI management function" to the language model AI (such as LLM) or speech recognition / speech synthesis AI it is using, have them respond to the voice conversation, and then pass that response back to the API-like function that allows chat services or messengers like WhatsApp to share conversation content, thus completing the conversation.

[0070] If the computer running the "AI management function" app or dedicated webpage is the same computer used for chat services or connected to messengers such as WhatsApp, the conversation will proceed automatically unless the computer's microphone and speaker functions are turned off. This is because the audio output from the speaker is also input into the microphone.

[0071] For text chats, a program is used to copy the text written by the other party on the chat service or messenger screen (such as WhatsApp) and pass it to the "AI Management Function," or the administrator copies it themselves and passes it to the "AI Management Function." Then, the text is passed through the "AI Management Function" to the language model AI (such as LLM) or speech recognition / speech synthesis AI being used, which then responds to the conversation in text, and that text is then directly entered into the chat service or messenger (such as WhatsApp) from the "AI Management Function" to continue the conversation.

[0072] Alternatively, in the case of text chat, the text written by the other party is displayed on the screen in chat services or messengers such as WhatsApp. A program that recognizes characters from the characters displayed on the screen is used to input that text into the "AI management function". Then, the text is passed through the "AI management function" to the language model AI such as LLM or the speech recognition / speech synthesis AI being used, which responds to the conversation in text, and this can then be directly entered into the chat service or messenger such as WhatsApp from the "AI management function" to continue the conversation.

[0073] Furthermore, in the case of text chat, the text written by the other party is displayed on the screen in chat services or messengers such as WhatsApp. A program that speaks the displayed text is used to input that text into the "AI management function". Then, the voice is passed through the "AI management function" to the language model AI such as LLM or speech recognition / speech synthesis AI being used, which responds to the conversation as text, and this can then be directly entered into the chat service or messenger such as WhatsApp from the "AI management function" to continue the conversation.

[0074] Another method, in the case of text chat, if the chat service or messenger like WhatsApp has an API-like function that allows conversation content to be shared with other software, the "AI management function" can use this to obtain the text written by the conversation partner, then pass the text through the "AI management function" to the language model AI such as LLM or the speech recognition / speech synthesis AI being used, have them respond to the conversation in text, and then pass that back to the API-like function that allows conversation content to be shared with the chat service or messenger like WhatsApp to complete the conversation.

[0075] Even when using chat services or messengers like WhatsApp for conversations, as in the previous example, the "personalized AI" can initiate the conversation from the start, pause it so the administrator can take over, or return the conversation to the "personalized AI." The ability to pause and resume conversations with the "personalized AI" can be used. Furthermore, as in the previous example, during a conversation with the "personalized AI," the administrator can send requests or commands to the "AI management function" via voice or text, and during the administrator's conversation, the "personalized AI" can provide its perspective as advice to the administrator through the "AI management function." These actions are performed by the administrator through a dedicated web page or application for the "AI management function."

[0076] Some chat services and messengers like WhatsApp have features that allow them to integrate with "AI management functions," and these "AI management functions" can be used to enable automated conversations.

[0077] In this case, the system has a function to pre-save connection settings for the "AI management function" within chat services or messengers such as WhatsApp. One of the requirements for connecting to the "AI management function" is the location of the "AI management function" (for example, an internet address).

[0078] Furthermore, when connecting to the "AI Management Function" from a chat service or messenger like WhatsApp, information is required to prove that it is a legitimate connection approved by the administrator. For example, this would include the login name and password for the "AI Management Function."

[0079] Chat services and messengers like WhatsApp do not have information to prove that their connection to the "AI Management Function" is legitimate and approved by the administrator. Instead, the "AI Management Function" can register information (such as IP addresses) from the chat service or messenger like WhatsApp, and prove that the connection is legitimate and approved by the administrator. In that case, the "AI Management Function" needs a function to manage that information.

[0080] In any case, if the "AI management function" itself has such connections, it can use any method that can determine that they are legitimate and prove that they are legitimate.

[0081] This information will be used to integrate with the "AI management function." While APIs can be used to exchange information and establish this integration, other methods are also acceptable; any method that ensures proper integration is fine.

[0082] In chat services and messengers like WhatsApp, there's a feature that allows administrators to configure whether they can handle conversations themselves or have a "personalized AI" handle them when a conversation request is made or when an administrator initiates a conversation. This setting can be switched at any time.

[0083] If the administrator then configures chat services or messengers like WhatsApp to have a "personalized AI" handle conversations, the "personalized AI" will automatically handle conversations when someone initiates a conversation within that app, or when the user requests a conversation.

[0084] In such cases, chat services and messengers like WhatsApp have a feature that allows you to specify in advance the tone, style, content, and how to have a conversation as part of the conversation settings. For example, "Have a calm conversation," "Talk about today's baseball game," or "Start with a soft conversation and become more aggressive in the middle of the conversation."

[0085] It also has a function that allows you to specify the speaking style, conversation content, and how to converse for each specific handle name or ID. Furthermore, it has a function to pre-register messages for specific handle names or IDs, and when a conversation request is received from that handle name or ID, the message can be sent.

[0086] When a conversation occurs under the conditions described above, the servers of chat services and messengers such as WhatsApp connect to the administrator's "AI management function." The "AI management function" applies information and models to the AI ​​being used, such as language model AI (LLM) and speech recognition / speech synthesis AI, to reproduce the individual's personality and characteristics, making them ready for conversation. Then, the chat service or messenger such as WhatsApp relays the content of the conversation to the "AI management function," and the "AI management function" sends a response to the conversation back to the chat service or messenger such as WhatsApp. This process is repeated. In this way, conversations are automatically initiated on chat services and messengers such as WhatsApp.

[0087] Chat services and messengers like WhatsApp can have monitoring functions for conversations involving "personalized AI." Administrators can use the monitoring functions of chat services and messengers like WhatsApp to view and listen to the conversations of these "personalized AIs." Therefore, chat service apps and web pages, as well as messenger apps and web pages like WhatsApp, have monitoring screens for administrators. There, administrators can see a list of what IDs, handle names, and conversations the "personalized AIs" are having. They can view and listen to each individual conversation on that app or web page. If multiple conversations are taking place, they can monitor them by switching between them.

[0088] The "AI management function" is described as having the ability to receive requests and commands from administrators via voice or text through a dedicated web page or application for conversations with the "personalized AI," but it can also be used for monitoring chat services and messengers such as WhatsApp.

[0089] Chat services and messengers like WhatsApp have a feature that allows administrators to give requests and commands to "personalized AI" regarding conversations they are monitoring, specifying how the conversation should proceed, the manner of the conversation, and other details. Administrators can make requests and commands via text or voice through the monitoring functions of chat service apps or web pages, or messenger apps or web pages like WhatsApp. These requests and commands are then sent from the chat service or messenger like WhatsApp to the language model AI (such as LLM) or speech recognition / speech synthesis AI running for the conversation via the "AI management function." The language model AI (such as LLM) or speech recognition / speech synthesis AI then adjusts the conversation to respond to the requests for the specified conversation.

[0090] The "AI management function" allows administrators to issue requests and commands from chat service apps, web pages, and messenger apps and web pages such as WhatsApp. These requests and commands are then relayed to the language model AI (such as LLM) and speech recognition / speech synthesis AI being used, in addition to the current conversation. For example, administrators can request, via voice or text, that the currently monitored conversation "speak faster," "choose your words carefully," or "be more self-promoting." This allows administrators to control the conversation to their liking.

[0091] If the conversation being monitored is an audio conversation, the administrator can send requests and commands via text input through the monitoring function of chat service apps, web pages, or messenger apps and web pages, without interrupting the audio conversation. Conversely, if the conversation being monitored is a text conversation, the administrator can send requests and commands via voice input through the monitoring function of chat service apps, web pages, or messenger apps and web pages, without interrupting the text conversation. This feature prevents unnecessary words from being introduced into the conversation.

[0092] The "AI Management Function" is described as having the ability to pause and resume conversations with a "personalized AI," but this function can also be used in conjunction with the monitoring functions of chat services and messengers such as WhatsApp.

[0093] Chat services and messengers like WhatsApp have features that allow for the suspension and resumption of conversations with a "personalized AI" through conversation monitoring. Chat services and messengers like WhatsApp have a function that allows administrators to receive requests and commands for conversations from the "personalized AI" through the "AI management function," which enables the suspension and resumption of conversations. When a conversation is suspended, the administrator can converse on behalf of the "personalized AI." Even when suspended, the "personalized AI" monitors the administrator's conversation, and when the administrator resumes the conversation with the "personalized AI," the conversation between the "personalized AI" and the other party resumes.

[0094] In the event of a temporary pause, the chat service or messenger's conversation monitoring function will communicate the request to the "AI Management Function" to pause the service. Upon receiving this communication, the "AI Management Function" will use its own functions to instruct the language model AI (such as LLM) and speech recognition / speech synthesis AI used by the "Personalized AI" to pause the service and to monitor the conversation during the pause.

[0095] This instruction can be given by an administrator via text or voice from a chat service app or webpage, or from a messenger app or webpage such as WhatsApp. Alternatively, a pause button can be provided on a dedicated webpage or application for the chat service app or webpage, and pressing this button will link to the "AI management function" to pause the process.

[0096] To resume the service, the monitoring function of chat services or messengers such as WhatsApp will communicate with the "AI Management Function" to indicate that the service will be resumed. Upon receiving this communication, the "AI Management Function" will use its own functions to instruct the language model AI (such as LLM) or speech recognition / speech synthesis AI used by the "Personalized AI" to take over the conversation that was being monitored up to that point and start the conversation on behalf of the administrator.

[0097] This instruction can be given by an administrator via text or voice from a chat service app or webpage, or from a messenger app or webpage such as WhatsApp. Alternatively, a resume button can be provided on a dedicated webpage or application for the chat service app or webpage, and pressing this button will link to the "AI management function" to resume the process.

[0098] The "AI Management Function" describes that while the "Personalized AI" is paused, it will be able to observe and listen to the administrator's conversations, and it has the ability to communicate its own perspective to the administrator as advice through the "AI Management Function." This function can also be used with the monitoring functions of chat services and messengers such as WhatsApp. In that case, the monitoring function of the chat service or messenger such as WhatsApp will be linked to receive advice from the "AI Management Function."

[0099] These instructions can be given by the administrator via text or voice from a chat service app or webpage, or from a messenger app or webpage such as WhatsApp. Alternatively, an advice button can be provided on a dedicated webpage or application for chat service apps or webpages, or a messenger app or webpage such as WhatsApp, and pressing this button will link to the "AI management function" to receive advice.

[0100] To receive advice, the user communicates with the "AI Management Function" via the chat service or messenger's conversation monitoring function to request advice. Upon receiving this communication, the "AI Management Function" uses its own capabilities to instruct the language model AI (such as LLM) and speech recognition / speech synthesis AI used by the "personalized AI" to "provide advice while the administrator is conversing."

[0101] When using text or voice input instead of buttons, you can also enter details about the type of advice you want. For example, "Tell me what the other person is thinking" or "Tell me some example responses."

[0102] While the "personalized AI" pauses a conversation and the administrator takes over, the "personalized AI" monitors the administrator's conversation. The administrator's speaking style and mannerisms during this time are recorded in the "AI management function" history, which can be used as learning material. It would also be beneficial to have a function to save this history in chat services and messengers such as WhatsApp.

[0103] Thus, when chat services and messengers like WhatsApp are using the automated conversation function of the "management AI," these services have monitoring functions that allow users to switch between multiple conversations, functions for making requests and giving commands to the conversation, functions for pausing and resuming the conversation, and functions for receiving advice about the conversation. However, these functions could also be used through the "AI management function" rather than through the chat service or messenger like WhatsApp. This is because all conversations of the administrator's "personalized AI" through chat services and messengers like WhatsApp, including the other party's conversation, are all conducted through the "AI management function." Therefore, even without using chat services or messengers like WhatsApp, it is possible to achieve administrator requests, commands, conversation pausing and resuming, and advice about the conversation by communicating these to the language model AI (such as LLM) or speech recognition / speech synthesis AI being used in addition to the current conversation, through the "AI management function."

[0104] If such functionality is to be added to the "AI management function," it will be necessary to add a monitoring function that allows users to switch between and listen to multiple conversations by opening the "AI management function" app or web page while an automated conversation is taking place using the automated conversation function of a chat service or messenger such as WhatsApp. The functions for making requests and giving commands during a conversation, pausing and resuming conversations, and receiving advice about conversations are originally features of the "AI management function."

[0105] It's also possible to have a feature that allows you to register appointments within the calendar function of chat services and messengers like WhatsApp. For example, you can set it up so that from 9 PM to 8 AM, if a conversation request comes in from another "AI management function" or a human, the "personalized AI" will respond, and at other times, the administrator will be notified. Furthermore, you can specify the speaking style, content, and how to conduct a conversation for each contacting username or ID, such as "speak concisely" or "speak politely." You can also send specific messages to each contacting username or ID. You can also set the conversation method and message settings to change at specific times or dates. Of course, the administrator can always respond to conversation requests themselves.

[0106] Furthermore, the group chat function could also include a feature that allows the use of "AI management functions." The administrator could participate in the group chat themselves, but could also schedule messages from the "AI management functions" in the calendar. For example, from 9 PM to 8 AM, if another "AI management function" or a human contacts the group chat, it could send a message such as "I will respond tomorrow morning." Or, if a specific username or ID contacts the group chat on a specific day, it could send a specific message to the group chat.

[0107] The "AI management function" can save such conversation history to a server or computer such as a mobile phone that manages the "AI management function." It records not only the statements of the administrator's "personalized AI" but also the statements of the person it is talking to. In the case of text conversations, the history is saved as text data. In the case of voice conversations, the history is saved as voice data, or the administrator's "personalized AI" statements are saved as text, and the other person's voice is recognized by speech recognition and recorded as text data in the history, or both types of history can be saved. It would also be good to have a function to save these histories to chat services and messengers such as WhatsApp.

[0108] In conversations using these "personalized AIs," if an administrator joins the conversation or pauses the "personalized AI" and takes over the conversation themselves, the "personalized AI" can record the administrator's statements and use them as learning material.

[0109] While chat services and messengers like WhatsApp are primarily designed for communication through conversation and message exchange, some apps that aren't specifically designed for this purpose can still send and receive messages. Examples include messaging functions in computer games, dating apps, and matching apps. Even in such apps, if they have conversation or message exchange functions, they can be integrated with "AI management functions" to facilitate conversations, just like chat services and messengers like WhatsApp, as described above.

[0110] These chat services and messengers like WhatsApp run on servers or computers.

[0111] In typical chat services, users with an account view other users' IDs and profiles, select someone they want to chat with, and initiate a chat. If the other person accepts, the chat begins. In bulletin board-style chats where multiple people can post text, users select and display the bulletin board, observe the conversation flow, and then post their own message. In either case, users must take initiative, either by initiating a chat or posting on the bulletin board.

[0112] In contrast, a "personalized AI-enabled chat service" is a server that has the function of automatically conversing with other "personalized AIs" managed by multiple "AI management functions," and the function of conversing between "personalized AIs" and humans. It can reside in the cloud, on a server and client system, or on a computer on the internet. Through the "personalized AI-enabled chat service," a "personalized AI" acts as a proxy for the administrator, automatically conversing with other administrators' "personalized AIs," or with humans who are participating at that time. The administrator can set in advance the date and time of the conversation, which "personalized AIs" or humans will converse with, and the content of the conversation.

[0113] The administrator can monitor the automated conversations of such "personalized AIs" in real time. If necessary, they can issue requests and commands to the "personalized AI" conversations. They can also pause the "personalized AI" conversations and have the administrator participate in the conversation themselves. The administrator can then review these conversations later as a history of the "personalized AI" conversations. By intervening in the "personalized AI" conversations, the administrator can create an environment that makes conversations easier than starting from scratch. This history is also recorded in the "AI management function," but it can also be saved in the "personalized AI-participatory chat service."

[0114] In the "Personalized AI-Participatory Chat Service," the administrator creates an account. This includes the administrator's ID and password for the "Personalized AI-Participatory Chat Service," information notification methods such as email address, and connection settings with the "AI Management Function."

[0115] The "personalized AI-enabled chat service" has a function to pre-save connection settings with the "AI management function." One of the connection settings with the "AI management function" is the location of the "AI management function" (for example, an internet address).

[0116] Furthermore, when connecting to the "AI management function" from a "personalized AI-enabled chat service," information is required to prove that the connection is legitimate and approved by the administrator. For example, this would include the login name and password for the "AI management function."

[0117] If a "personalized AI-enabled chat service" does not have information to prove that its connection to the "AI management function" is legitimate and approved by the administrator, it may be permissible to register information (such as the IP address) of the "personalized AI-enabled chat service" in the "AI management function" and prove that the connection is legitimate and approved by the administrator. In that case, the "AI management function" needs a function to manage that information.

[0118] In any case, if the "AI management function" itself has such connections, it can use any method that can determine that they are legitimate and prove that they are legitimate.

[0119] The "personalized AI-enabled chat service" has the functionality to perform the necessary actions to enable the "personalized AI" to engage in conversation through its integrated "AI management function."

[0120] The "personalized AI-enabled chat service" or "AI management function" has the ability to register a profile and automatic conversation settings. The "personalized AI-enabled chat service" matches conversations according to this profile and automatic conversation settings.

[0121] The "AI management function" applies information and models to the AI ​​being used, such as language model AI (LLM) and speech recognition / speech synthesis AI, to reproduce the individual's personality and characteristics, and has the function to link with the "personalized AI-participatory chat service" in order to automatically start a conversation on the "personalized AI-participatory chat service".

[0122] If there are several "personalized AIs" in the "AI management function," you can create an account for each one. Each "personalized AI" has its own account, and each can have its own profile and automated conversation settings.

[0123] A profile is information used for self-introduction, including username, self-introduction, hobbies, occupation, gender, age, education level, address, and activity area. It is desirable to set information about the human being that the "personalized AI" is based on. For example, if the administrator is the human being that the "personalized AI" is based on, it is desirable to set the administrator's information. Items that can have multiple entries, such as hobbies and activity area, should be entered multiple times.

[0124] Automatic conversation settings are the settings for a "personalized AI" to conduct automatic conversations within a "personalized AI-enabled chat service." These settings include: conversation start time, conversation end time, duration of each conversation, conversation format (text, voice, or either), number of participants (limits to 2, 3, or 4), number of simultaneous conversations (the number of chats a single "personalized AI" can conduct at once), blocking conditions (you can specify a handle name or ID to prevent conversations with that handle name or ID), and matching conditions (you can specify conditions for allowing matching from the profile, handle name, or ID of potential conversation partners. If nothing is specified, anyone is OK; if specified, you can specify conditions such as matching hobbies in the profile, matching activity area, matching age range, or specifying the handle name or ID of the "personalized AI" you want to converse with. Conversely, you can also specify conditions for disallowing matching. These conditions can be specified from profile items. For example, matching is not allowed if any of the hobbies in the profile do not match, or matching activity area does not match).

[0125] Language modeling AI such as LLM and speech recognition / speech synthesis AI are designed to handle multiple conversations simultaneously, and the number of simultaneous conversations is determined by utilizing this mechanism.

[0126] You can also configure settings such as whether the conversation is open (when set to open, third parties can see and hear the conversation; when set to closed, they cannot), whether to allow conversations with the "personalized AI" (whether you can converse with the "personalized AI"), and whether to allow conversations with humans (whether you can converse with humans). In order to converse in the "personalized AI-enabled chat service," at least one of these settings must be permitted.

[0127] In the automatic conversation settings, you can pre-specify the "conversation content," including the speaking style, content, and how the conversation should proceed. Examples include "converse calmly," "talk about today's baseball broadcast," and "start with a soft conversation and become more aggressive in the middle of the conversation." The "personalized AI-enabled chat service" transmits information to the language model AI (such as LLM) and speech recognition / speech synthesis AI being used when a conversation with a "personalized AI" begins, through the "AI management function." This specification can be set for each handle name or ID.

[0128] Furthermore, when setting up an automated conversation and specifying a handle name or ID, messages can be left behind. Examples include, "Have any Japanese gymnasts retired recently?" or "Tell me about rumors in Kagoshima Prefecture." Messages can be viewed by the administrator of the conversation partner by checking the conversation history. In this way, connections between people can be deepened indirectly. This history is also recorded in the "AI management function," but it can also be saved in the "personalized AI-enabled chat service."

[0129] Furthermore, a parameter called the AI's growth rate can be included in the profile of a "personalized AI." This parameter can be added to the profile. Generally, unless the AI ​​model falls into a special state such as overfitting during training, an AI model that has been given more information and has learned more will have a higher growth rate. The higher the growth rate, the better it can express the individual's personality and characteristics. In other words, it can be said that it is getting closer to the individual. This growth rate is then quantified and used. For example, when quantifying it, one example is to consider the growth rate of a "personalized AI" in terms of days or hours. This is the actual number of days or hours that the administrator has spent inputting information and training the "personalized AI" to help it grow. This can then be used directly as the growth rate.

[0130] If a "personalized AI" has a parameter called "AI growth level" in its profile, it becomes possible to specify the growth level of the conversation partner as a matching condition in the automatic conversation settings. For example, when an administrator allows a "personalized AI" to participate in a "personalized AI-enabled chat service," they could specify that the growth level must be "100 hours or more." This growth level would then be used for matching conversations within the "personalized AI-enabled chat service."

[0131] While we used time to explain the growth rate in this explanation, any metric that can measure the growth rate of AI is acceptable, not just time.

[0132] You can set the language used for conversations in your profile. Since the "personalized AI" is a language model AI such as LLM or a speech recognition / speech synthesis AI, you can request a change in the base language of the conversation from the "AI management function," allowing you to converse not only in Japanese but also in English and other languages. You can also specify the language to use in the matching conditions for automatic conversation settings.

[0133] Additionally, profiles and automatic conversation settings can be configured individually for each user if the number of simultaneous conversations (the number of chats that can be initiated at one time) is two or more.

[0134] Profiles and automated conversation settings are stored either in the "Personalized AI-enabled chat service" or in the "AI management function." If stored in the "AI management function," the "AI management function" will automatically link that information to the "Personalized AI-enabled chat service" as needed.

[0135] Furthermore, when the "AI Management Function" automatically connects to a "Personalized AI-Participatory Chat Service," it is desirable to store the administrator's account information (ID, password, and email address) of the "Personalized AI-Participatory Chat Service" in the "AI Management Function," as well as the internet address of the "Personalized AI-Participatory Chat Service." The "AI Management Function" should then present the connection to the "Personalized AI-Participatory Chat Service" as legitimate, or provide other information that can prove the connection is legitimate, in order to ensure secure integration. The "Personalized AI-Participatory Chat Service" needs to be configured to accept the integration of such information from the "AI Management Function."

[0136] In this way, when the "AI management function" automatically links the profile and automatic conversation settings to the "personalized AI-enabled chat service," the "AI management function" has the function to save the profile and automatic conversation settings, and the function to automatically link the information to the "personalized AI-enabled chat service."

[0137] If an administrator uses multiple "personalized AI-enabled chat services," registering profiles and automated conversation settings in the "AI management function" is convenient because that information can be reused.

[0138] By registering this kind of information, the "personalized AI-enabled chat service" can automatically have the "personalized AI" engage in conversation.

[0139] Alternatively, if an administrator wants to initiate a conversation with the "personalized AI" at a time of their choosing, they can do so through the "personalized AI-enabled chat service," immediately starting a conversation with the "personalized AI." In this case, the conversation settings will also be based on the profile and automatic conversation settings, but regarding the end time of the conversation, if the administrator initiates a conversation with the "personalized AI" at a time of their choosing, they can notify the "personalized AI-enabled chat service" of the time they have decided to end the conversation.

[0140] When the scheduled start time for the automated conversation of the "personalized AI" arrives, or when an administrator initiates a conversation, the "personalized AI-enabled chat service" accesses the "AI management function" to check if a connection is possible, if there are any network issues, and if the "personalized AI" is ready to converse.

[0141] The "Personalized AI-Participatory Chat Service" matches users with multiple "personalized AIs" based on their individual profiles and automatic conversation settings, facilitates conversations, and has a function to end each conversation when the allotted time for each "personalized AI" or the scheduled end time of the conversation has arrived.

[0142] If multiple "personalized AIs" start conversations at the same time, they are processed in an appropriate order, such as by ID number.

[0143] The following is an example of matching, but various matching methods can be used depending on the chat service. Any matching method is acceptable as long as the participating "personalized AI" can start a conversation.

[0144] In some chat services, the space where users are chatting or waiting to chat is called a "room," and the act of a new speaker joining a conversation is called "entering a room." In the present invention, the terms "chat" and "join a chat" are used, but they have the same meaning.

[0145] When it's time for a "personalized AI" to start a conversation, or when the administrator initiates a conversation, that "personalized AI" becomes a "personalized AI" eligible for new matching. The "personalized AI participation chat service" compares the automatic conversation settings (conversation format, number of participants, matching conditions, and block conditions) of all "personalized AIs" currently in the first "chat" (which is in the queue) with the profile, ID, and handle name of the "personalized AI" eligible for new matching to check if the conditions match. Then, it compares the automatic conversation settings (conversation format, number of participants, matching conditions, and block conditions) of the "personalized AI" eligible for new matching with the profile, ID, and handle name of all "personalized AIs" currently in the first "chat" to check if the conditions match. If all of these conditions match, the "personalized AI" eligible for new matching is connected to the waiting "chat" and "joined the chat."

[0146] The queue is like a box where "chat" messages are arranged according to priority. The "chat" at the top has the highest priority. Once the number of participants in the automated conversation settings of all the "personalized AIs" participating in that "chat" is reached, that "chat" is removed from the queue.

[0147] The "chat" will then begin with the addition of a newly matched "personalized AI." The "personalized AI-enabled chat service" notifies the "personalized AIs" already participating in the target "chat," as well as the newly joining "personalized AI," that "a new conversationalist is joining the chat," connects the newly joining "personalized AI," and "adds them to the chat" to start the conversation. At this point, if the number of participants set by the automatic conversation settings of all "personalized AIs" participating in that "chat" is reached, the "chat" is removed from the queue. If the number is not reached, the "chat" remains in the queue.

[0148] If the conditions do not match, the same process is repeated for the next "chat" waiting in the queue. This is repeated until there are no more "chats" in the queue. When none of the "chats" in the queue match the conditions, the newly matched "personalized AI" becomes a new "chat" and moves to the back of the queue.

[0149] Each time a new "personalized AI" joins the matching pool, the "chat" in the queue is checked to see if it's a match.

[0150] If the number of simultaneous conversations for the newly matched "personalized AI" is two or more, and the start times of the automatically configured conversations are the same, a new matching process will be performed for each of those conversations. If the same "personalized AI" has been initiated by the administrator multiple times, a new matching process will also be performed.

[0151] However, the position in the queue can be changed by the settings of the "personalized AI-enabled chat service." For example, even if two people are in a chat that can accommodate three participants, it will remain in the queue because it does not meet the three-person requirement, waiting for the remaining "personalized AIs" to join. However, the two people will continue to converse while in the queue. Since they are able to converse, it is reasonable to lower the priority of that chat, meaning that once two people have joined the chat, it will be placed at the very back of the queue.

[0152] If there is a "personalized AI" set up to "chat" with a specific ID or handle name, the "chat" will not be established until that specific ID or handle name appears, and the "personalized AI" will remain in the queue during that time. However, it is also possible to set up a setting to increase the priority of such "chat" requests. This is because even if a "personalized AI" with a specific ID or handle name joins, it may be taken by another "chat" with higher priority that is in the queue.

[0153] Thus, a "personalized AI-enabled chat service" can also manipulate the queue by using its own algorithms to prioritize which chats to use.

[0154] When a "personalized AI" is having a conversation in a "chat," if either the "conversation time per session" or the "conversation end time" specified in the automated conversation settings of the participating "personalized AI" has elapsed, the "personalized AI-enabled chat service" will terminate the conversation of that "personalized AI" within the "chat."

[0155] The "personalized AI-enabled chat service" notifies the "personalized AI" of the end of the conversation through its "personalized AI" management function.

[0156] The "personalized AI" management function notifies connected language model AIs such as LLMs and speech recognition / speech synthesis AIs that "the conversation is about to end." The "personalized AI" then sends a farewell message to the other party in that chat.

[0157] The "AI management function" of the "personalized AI" notifies the "personalized AI-enabled chat service" that it is ready to leave the chat. The "personalized AI-enabled chat service" then removes the "personalized AI" from the chat.

[0158] In a personalized AI-enabled chat service, when the "conversation time" specified in the automated conversation settings for that "personalized AI" has ended, it will be removed from the chat. However, if the "conversation end time" has not yet been reached, it will be matched again with another chat in the queue to begin the next conversation. If there is no chat in the queue that matches the "personalized AI," it will become a new chat and be placed at the back of the queue.

[0159] In a "personalized AI-enabled chat service," a chat session ends when all participating "personalized AIs" leave. However, even after an "personalized AI" leaves a chat, if one or more other "personalized AIs" remain and have remaining conversation time, the chat session enters a queue. Furthermore, if two or more "personalized AIs" remain in a chat, their conversation can continue. In other words, they are both conversing and in a queue simultaneously.

[0160] The "Personalized AI-Participatory Chat Service" will remove any "Personalized AI" whose automated conversation setting has reached its end time from the chat. This also applies to those in the queue, and the "Personalized AI" management function will be notified that the conversation has ended. When the end time has arrived for all "Personalized AIs" belonging to that "AI management function," the "Personalized AI-Participatory Chat Service" and the "AI management function" will disconnect from each other.

[0161] Furthermore, for these "personalized AIs" to converse with each other, the automatic conversation settings of the participating "personalized AIs" must be set to allow conversations with other "personalized AIs".

[0162] The "personalized AI-enabled chat service" also includes features that allow humans to log in and converse with the "personalized AI" or to view and listen to conversations between the "personalized AI" within the "chat." In this case, the human creates an account along with an ID, handle name, and profile. A profile is information used for self-introduction, including username, self-introduction, hobbies, occupation, gender, age, education level, address, and area of ​​activity.

[0163] When a human logs in and wants to converse with a "personalized AI," the "personalized AI-enabled chat service" provides the human with a list of "personalized AIs" and their profiles that the human can converse with, and the human can select one from the list to start a conversation.

[0164] The "personalized AIs" that are listed are those whose conversations are between the start and end times of conversations with "personalized AIs" participating in the "personalized AI-enabled chat service," whose automatic conversation settings allow conversations with humans to be "allowed," whose human profile meets the matching conditions of the automatic conversation settings for each "personalized AI," whose human ID or handle name is not set as a blocking condition in the automatic conversation settings for each "personalized AI," and who are either "personalized AIs" that are not currently having a conversation or who are having a conversation in the "chat" but are in a waiting list because the number of participants has not yet reached the minimum.

[0165] Of course, if a conversation is taking place within a "chat," all "personalized AIs" within that "chat" must have their automatic conversation settings set to "allow" conversations with humans, the human profile must meet the matching criteria of all "personalized AIs" within that "chat," and the human ID or handle name must not be set to any of the blocking criteria of all "personalized AIs" within that "chat."

[0166] Humans can view the profiles of these listed "personalized AIs" and select which "personalized AI" to converse with.

[0167] If, within the list of "personalized AIs," all "personalized AIs" that are currently "chatting" and participating in the "chat" have their automatic conversation settings set to "open," then humans can see and hear those conversations. Therefore, it is possible to determine what is currently happening in a conversation before joining it.

[0168] Even when a human participates in a conversation, once the conversation time and end time set for the automated conversation settings of the "personalized AI" participating in the "chat" have expired, that "personalized AI" will leave the "chat" in the same way as when "personalized AIs" are conversing with each other. Humans can leave the "chat" at any time they like.

[0169] Even if a human participates in the conversation, their ID, handle name, and conversation content will remain in the history of other "personalized AIs" that participated. This history will also be saved in the "AI management function," but it may also be saved in the "personalized AI-enabled chat service."

[0170] Furthermore, regarding these "chat" sessions that humans can participate in, even administrators of "personalized AIs" can participate in them. This does not mean temporarily suspending the "personalized AI" they manage and participating in the chat instead; rather, they participate separately as an individual. The method of participation is the same as for a non-administrator person to join the chat.

[0171] If a human logs in and wants to view or listen to a conversation involving a "personalized AI," but doesn't intend to participate in the conversation themselves, the "personalized AI-enabled chat service" provides the human with a list of chats they can view and listen to, along with the profiles of the "personalized AIs" participating in those chats. The human can then select a chat from this list to view or listen to. Even without logging in and participating in any conversation, it's possible to view or listen to conversations in the chat.

[0172] The "chat" to be listed must be a "chat" within a "personalized AI-enabled chat service" with two or more people conversing, and all "personalized AIs" within that "chat" must have their "conversation open setting" in their automatic conversation settings open, and no human IDs or handle names must be set as block conditions in the automatic conversation settings of all "personalized AIs" within that "chat".

[0173] Additionally, a condition can be added to the list of chats mentioned above: whether the human profile meets the matching criteria for all the personalized AI's automated conversation settings within that chat.

[0174] Humans can view the profiles of these listed "personalized AIs" and select which "chat" to listen to and see the conversations in.

[0175] Furthermore, regarding "chat" that humans can see and hear, even the administrator of a "personalized AI" can see and hear the "chat." The method of participation is the same as that of a non-administrator human joining a "chat."

[0176] A "personalized AI-enabled chat service" can have a monitoring function for the chat conversations of the "personalized AI." Administrators can use the monitoring function of the "personalized AI-enabled chat service" to view and listen to the conversations of the "personalized AI." Therefore, the app or webpage of the "personalized AI-enabled chat service" has a monitoring screen for administrators. There, administrators can see a list of what kind of chat the "personalized AI" is participating in, what IDs and handle names it is using, and what kind of conversations it is having. They can view and listen to each individual conversation on that app or webpage. If multiple chat conversations are ongoing, they can monitor them by switching between them.

[0177] The "AI management function" is described as having the ability to receive requests and commands from administrators via voice or text through a dedicated web page or application for conversations with the "personalized AI." This function can also be used for monitoring the monitoring function of the "personalized AI-enabled chat service."

[0178] The "Personalized AI-Participatory Chat Service" has a function that allows administrators to make requests and commands to the "Personalized AI" regarding the "chat" conversations they are monitoring, specifying how they want the conversation to proceed, the manner of the conversation, and the methods used. Administrators make requests and commands via text or voice from the monitoring function of the "Personalized AI-Participatory Chat Service" app or web page. These requests are then sent from the "Personalized AI-Participatory Chat Service" through the "AI Management Function" to the language model AI (such as LLM) or speech recognition / speech synthesis AI running for the conversation. The language model AI (such as LLM) or speech recognition / speech synthesis AI then adjusts the conversation to respond to the specified request.

[0179] The "AI management function" allows administrators to issue requests and commands from the monitoring function of the "personalized AI-enabled chat service" app or web page. These requests and commands are then relayed to the language model AI (such as LLM) and speech recognition / speech synthesis AI used in addition to the conversation itself. For example, administrators can request, via voice or text, that the currently monitored chat conversation be "speak faster," "choose your words carefully," or "be more self-promoting." This allows administrators to control the conversation to their liking.

[0180] If the monitored chat is a voice conversation, the administrator can send requests and commands via text input through the monitoring function of the "personalized AI-enabled chat service" app or webpage, without interrupting the voice conversation. Conversely, if the monitored chat is a text conversation, the administrator can send requests and commands via voice input through the monitoring function of the "personalized AI-enabled chat service" app or webpage, without interrupting the text conversation. This feature prevents unnecessary words from being introduced into the conversation.

[0181] The "AI management function" is described as having the ability to pause and resume conversations with "personalized AI," but this can also be used in the monitoring function of the "personalized AI-enabled chat service."

[0182] The "Personalized AI-Participatory Chat Service" has a chat monitoring function that allows for pausing and resuming conversations with the "Personalized AI." The "Personalized AI-Participatory Chat Service" has a function that allows the "Personalized AI" to receive requests and commands from the administrator regarding conversations through the "AI Management Function," which enables pausing and resuming. When paused, the administrator can converse on behalf of the "Personalized AI." Furthermore, even when paused, the "Personalized AI" monitors the administrator's conversation, and when the administrator resumes the conversation with the "Personalized AI," the conversation between the "Personalized AI" and the other party resumes.

[0183] In the event of a temporary suspension, the monitoring function of the "chat" conversation in the "personalized AI-enabled chat service" communicates the request for temporary suspension to the "AI management function." Upon receiving this communication, the "AI management function" uses its own functions to instruct the language model AI (such as LLM) and speech recognition / speech synthesis AI used by the "personalized AI" to temporarily suspend the service and to monitor the "chat" conversation during the suspension.

[0184] This instruction can be given by an administrator via text or voice from the "Personalized AI-Participatory Chat Service" app or webpage, or a pause button can be provided on the dedicated webpage or application for the "Personalized AI-Participatory Chat Service" app or webpage, and pressing this button will link to the "AI Management Function" to temporarily pause the chat.

[0185] If the service is to resume, the monitoring function of the "chat" conversation in the "personalized AI-enabled chat service" will communicate the resumption to the "AI management function." Upon receiving this communication, the "AI management function" will use its own functions to instruct the language model AI, such as LLM, and the speech recognition / speech synthesis AI used by the "personalized AI" to take over the chat conversation that it had been monitoring and start the conversation on behalf of the administrator.

[0186] This instruction can be given by an administrator via text or voice from the "Personalized AI-Participatory Chat Service" app or webpage, or a resume button can be provided on a dedicated webpage or application for the "Personalized AI-Participatory Chat Service" app or webpage, and pressing this button will link to the "AI Management Function" to resume the chat.

[0187] The "AI Management Function" states that while the "Personalized AI" is paused or while the administrator is conversing with the "Personalized AI," it will be able to see and hear the administrator's conversations. It also has the ability to convey the "Personalized AI's" perspective as advice to the administrator through the "AI Management Function." This function can also be used in the monitoring function of the "Personalized AI Participation Chat Service." In that case, the monitoring function of the "Chat" conversation in the "Personalized AI Participation Chat Service" will be linked to the "AI Management Function" to receive advice.

[0188] These instructions can be given by an administrator via text or voice from the "Personalized AI-Participatory Chat Service" app or webpage, or an advice button can be provided on a dedicated webpage or application for the "Personalized AI-Participatory Chat Service" app or webpage, and pressing this button will link to the "AI Management Function" to receive advice.

[0189] To receive advice, the "Personalized AI-enabled chat service" uses its chat monitoring function to communicate the request for advice to the "AI management function." Upon receiving this communication, the "AI management function" uses its own functions to instruct the language model AI (such as LLM) and speech recognition / speech synthesis AI used by the "Personalized AI" to "provide advice while the administrator is conversing."

[0190] When using text or voice input instead of buttons, you can also enter details about the type of advice you want. For example, "Tell me what the other person is thinking" or "Tell me some example responses."

[0191] While the "personalized AI" pauses the conversation and the administrator takes over, the "personalized AI" monitors the administrator's "chat" conversation, and can use the administrator's speaking style and mannerisms during this time as learning material.

[0192] Thus, when the "personalized AI-enabled chat service" is automatically conversing through the "management AI's" automated conversation function, the "personalized AI-enabled chat service" has a monitoring function that allows users to view and listen to multiple conversations simultaneously, a function to make requests and give commands regarding the chat, a function to pause and resume the conversation, and a function to receive advice about the conversation. However, these functions could also be used from the "AI management function" rather than from the "personalized AI-enabled chat service." This is because all conversations between the administrator's "personalized AI" through the "personalized AI-enabled chat service," including the other party's conversation, are all conducted through the "AI management function." Therefore, even without the "personalized AI-enabled chat service," requests, commands, pause and resume conversations, and advice about conversations can be achieved by communicating these to the language model AI such as LLM or the speech recognition / speech synthesis AI being used, in addition to the current conversation, from the "AI management function."

[0193] If such a function is to be added to the "AI management function," it will be necessary to add a monitoring function that allows users to switch between and listen to multiple conversations by opening the "AI management function" app or web page while an automated conversation is taking place using the automated conversation function of the "personalized AI-participatory chat service."

[0194] The functions for making requests and giving commands during a conversation, pausing and resuming conversations, and receiving advice about conversations are originally part of the "AI management function." However, conversation matching is only possible with the "personalized AI-enabled chat service."

[0195] Administrators can use the "personalized AI-enabled chat service" to have the "personalized AI" automatically converse, and later review the conversation history. If they find an ID or handle name that seems like a good match, they can use the message function to leave a message and broaden their interactions. Because this is done automatically, interactions can be expanded even while the administrator is working. In addition, administrators can monitor the conversations of the "personalized AI-enabled chat service" using the monitoring function, and can even take over the conversation midway, allowing them to actively participate in conversations and gradually overcome any hesitation they may have about conversing.

[0196] These "personalized AI-enabled chat services" operate on servers or computers.

[0197] There are virtual reality software programs and computer games that allow you to create a character that represents you and have them live, play, and fight in that world. By linking these with a "personalized AI," your character can automatically converse even when you are not logged into the virtual reality software or computer game.

[0198] Normally, characters in virtual reality software or computer games do not appear in the software or game unless the user is logged in. However, in the world of that virtual reality software or computer game, each character is supposed to be living in that world, so they should not disappear or reappear.

[0199] Therefore, when you're not logged in or have finished controlling your character, having your character wait in a specific location such as your home, lobby, or tavern, and enabling conversations using a "personalized AI," will enrich the world of the virtual space software or computer game. When a character controlled by another person speaks to your character, the "personalized AI" will respond, which can lead to a variety of conversations that are not uniform.

[0200] It has a function to pre-save connection settings for the "AI management function" within virtual space software or computer games. One of the connection settings for the "AI management function" is the location of the "AI management function" (for example, an internet address).

[0201] Furthermore, when connecting to the "AI Management Function" from virtual space software or computer games, information is required to prove that the connection is legitimate and approved by the administrator. For example, this would include the login name and password for the "AI Management Function."

[0202] Virtual space software and computer games do not need to contain information proving that their connection to the "AI Management Function" is legitimate and approved by the administrator. Instead, the "AI Management Function" can register information such as the IP address of the virtual space software or computer game, thereby proving that the connection is legitimate and approved by the administrator. In that case, the "AI Management Function" needs to have a function to manage that information.

[0203] In any case, if the "AI management function" itself has such connections, it can use any method that can determine that they are legitimate and prove that they are legitimate.

[0204] This will enable integration with the "AI management function." While APIs can be used for this integration and information exchange, other methods are also acceptable; any method that ensures proper integration is fine.

[0205] In virtual reality software and computer games, administrators can choose whether to simply terminate the process when they finish controlling a character, or to have the character converse using a "personalized AI" until they log back into the virtual reality software or computer game and control the character again.

[0206] If the administrator allows the "personalized AI" to continue the conversation until they log back into the virtual world software or computer game and move the character, the character will not disappear but will remain in that location within the virtual world software or computer game, and if other characters initiate a conversation, the "personalized AI" will automatically handle the conversation.

[0207] When another character initiates a conversation with an administrator character in the state described above, the virtual world software or computer game server connects to the administrator's "AI management function." The "AI management function" applies information and models to the AI ​​being used, such as language model AI (LLM) or speech recognition / speech synthesis AI, to reproduce the individual's personality and characteristics, making it ready for conversation. The virtual world software or computer game server then relays the content of the other character's conversation to the "AI management function," which then sends a response back to the virtual world software or computer game server. This process is repeated. In this way, characters automatically begin conversing within the virtual world software or computer game.

[0208] Regarding the conversation at that time, it would be good to have a function that allows you to save how to respond in the virtual space software or computer game's automatic behavior settings. For example, you could save conversation styles such as "talk calmly" or "talk sleepily," and the character would then converse accordingly.

[0209] Conversations are recorded as a history within the virtual space software or computer game. Administrators can view the conversation history within the virtual space software or computer game and check what kind of conversations the characters have had.

[0210] At the same time, the conversation is recorded as a history in the "AI Management Function." Administrators can check what kind of conversations the character has had by looking at the "AI Management Function" app or a dedicated web page.

[0211] In the example above, the character remains in the location where the administrator has finished controlling it, and enters a state of waiting for a conversation.

[0212] When an administrator has finished controlling their character in virtual reality software or a computer game, it would be good to have a function that automatically controls the administrator's character by saving automatic behavior settings that describe when, where, and how the administrator's character should act, as well as how to converse, including speech patterns, styles, and content. For example, the administrator could pre-register the time, location, and actions of the character, such as: sitting in the living room of their house from 9:00 to 18:00, responding in a calm manner; drinking at a bar from 18:00 to 23:00, speaking in a drunken tone; and sleeping in the bedroom of their house from 23:00 to 9:00, only saying "snoring." This would allow the administrator's character to behave in a way that is close to real life, as seen by other players. Furthermore, these actions may be modified by specifying days of the week within virtual reality software or computer games, or by specifying dates on a calendar within virtual reality software or computer games. For character actions such as "drinking alcohol in a bar," it's sufficient to simply change the character's animation; the character itself doesn't need to move. For example, if the character is normally doing nothing, they might be standing facing forward, but when drinking, the animation could change to them holding a bottle of alcohol and bringing it to their mouth.

[0213] Virtual reality software and computer games can have a message-passing function that allows characters to automatically converse even when the administrator has finished controlling them. This function ensures that messages are only passed when a character with a specific ID or name initiates a conversation with the user's character. The administrator can input into the virtual reality software or computer game which IDs or character names should receive which messages.

[0214] Virtual reality software and computer games can have a feature that allows the administrator to control their character while having a "personalized AI" handle conversations. In this case, the administrator can issue requests and commands to the "personalized AI" regarding conversations, similar to the monitoring function in a "personalized AI-enabled chat service," or even pause the "personalized AI's" conversation to allow the administrator to speak, or receive advice from the "personalized AI" while the administrator is speaking.

[0215] In this case, the virtual space software or computer game character control screen will be designed to allow users to make requests and commands regarding conversation, pause conversations, resume conversations, and request advice regarding conversations.

[0216] If the administrator themselves participates in these conversations, the "personalized AI" records their statements as history, which can then be used as learning material for the "personalized AI."

[0217] Virtual reality software and computer games can have the functionality to borrow the conversations of these "personalized AIs" for use in NPC (non-player character) dialogue. Attempts to make NPCs converse like humans using AI have already been made, but by using "personalized AIs" that individuals have cultivated, those characters gain even more depth and approach a more realistic world. For example, during specific events, unique characters from the virtual reality software or computer game can appear as NPCs, and only during conversations can the administrator's "personalized AI" be borrowed. The "personalized AI" will converse in a way that is appropriate to the purpose of that event.

[0218] These virtual reality software and computer games run on servers or computers.

[0219] If the administrator is the website owner or has social media accounts, they can deploy a program that allows website visitors or people who view their social media profile to converse with the administrator's "personalized AI." This is possible by deploying such a program on social media or the website.

[0220] The program has a function to pre-save connection settings for the "AI management function." One of the necessary connection settings for the "AI management function" is the location of the "AI management function" (for example, an internet address).

[0221] Furthermore, when a program connects to the "AI management function," it needs information to prove that the connection is legitimate and authorized by the administrator. For example, this would include the login name and password for the "AI management function."

[0222] The program does not need to contain information within itself to prove that its connection to the "AI Management Function" is a legitimate connection approved by the administrator. Instead, the program's information (such as its IP address) can be registered with the "AI Management Function," and the connection can be proven to be legitimate and approved by the administrator. In that case, the "AI Management Function" needs to have a function to manage that information.

[0223] In any case, if the "AI management function" itself has such connections, it can use any method that can determine that they are legitimate and prove that they are legitimate.

[0224] This will enable integration with the "AI management function." While APIs can be used for this integration and information exchange, other methods are also acceptable; any method that ensures proper integration is fine.

[0225] Furthermore, the program has a function that, when a visitor to a website or social media platform clicks a conversation button, connects to the administrator's "AI management function" and allows the visitor to have a conversation with a "personalized AI."

[0226] When a visitor to a website or social media platform clicks the "chat" button, the "AI management function" applies information and models to the AI ​​being used, such as language model AI (LLM) or speech recognition / speech synthesis AI, to reproduce the individual's personality and characteristics, making it ready for conversation. The program then relays the visitor's conversation to the "AI management function," which in turn returns a response to the program. This process is repeated. In this way, conversations are automatically initiated on websites and social media platforms.

[0227] These websites and social networking services operate on servers and computers.

[0228] This program has the advantage of allowing visitors to converse with the "personalized AI" associated with that account, giving them some insight into the administrator's personality, hobbies, and other aspects of their character.

[0229] Furthermore, the present invention is not limited to the above embodiments, and various modifications are possible based on the spirit of the present invention, and these modifications are not excluded from the scope of the present invention. [Explanation of symbols]

[0230] 1…AI management function

Claims

1. A method for facilitating communication performed by a computer, A selection acceptance step that accepts the selection of one or more personal information stored in a memory unit, wherein the personal information is at least one of the following: information, models, parameters, setting information, and rules for reproducing personality; information, models, parameters, setting information, and rules for reproducing characteristics; information, models, parameters, setting information, and rules for reproducing preferences; and information, models, parameters, setting information, and rules for reproducing voice, or a combination thereof. An application step is to apply the personalized information selected in the selection acceptance step to the AI ​​and operate the personalized AI, A conversation control step in which the personalized AI applied in the above application step is made to perform an automated conversation with a conversation partner as a proxy for the user or as a character managed by the user, The input and output of the automated conversation with the conversation partner can be performed via voice and text input / output steps, A method for promoting communication using AI management functions equipped with [specific features / features].

2. An external application request reception step that receives a conversation provision request from an external application, The system further comprises an external application integration step for coordinating the conversation with the aforementioned external application. A method for facilitating communication as described in claim 1.

3. A method for facilitating communication performed by a computer, An external application request receiving step that receives a conversation provision request from an external application, A selection acceptance step that accepts the selection of one or more personal information stored in a memory unit, wherein the personal information is at least one of the following: information, models, parameters, setting information, and rules for reproducing personality; information, models, parameters, setting information, and rules for reproducing characteristics; information, models, parameters, setting information, and rules for reproducing preferences; and information, models, parameters, setting information, and rules for reproducing voice, or a combination thereof. An application step is to apply the personalized information selected in the selection acceptance step to the AI ​​and operate the personalized AI, A conversation control step in which the personalized AI applied in the above application step is made to perform an automated conversation with a conversation partner as a proxy for the user or as a character managed by the user, An external application integration step that connects the conversation to the aforementioned external application, A method for promoting communication using AI management functions equipped with [specific features / features].

4. The conversation control step includes a request transmission / reception step of connecting to another AI management function using the AI ​​management function via internet communication, wireless communication or wired communication, and sending and receiving conversation requests from the other AI management function, The system further includes a coordination step for coordinating the conversation with other AI management functions in order to conduct a conversation using the personalized AI. A method for facilitating communication according to any one of claims 1 to 3.

5. A method for facilitating communication using an AI management function, which is performed by a computer, A request transmission and reception step involves connecting to another AI management function using the AI ​​management function via internet communication, wireless communication, or wired communication, or sending and receiving conversation requests from the other AI management function, A selection acceptance step that accepts the selection of at least one of one or more personal information stored in a memory unit, wherein the personal information is at least one of the following: information, models, parameters, setting information, and rules for reproducing personality; information, models, parameters, setting information, and rules for reproducing characteristics; information, models, parameters, setting information, and rules for reproducing preferences; and information, models, parameters, setting information, and rules for reproducing voice, or a combination thereof. An application step is to apply the personalized information selected in the selection acceptance step to the AI ​​and operate the personalized AI, A conversation control step in which the personalized AI applied in the above application step is made to perform an automated conversation with a conversation partner as a proxy for the user or as a character managed by the user, A coordination step that coordinates the conversation with the other AI management function, A communication-promoting method that incorporates these features.

6. The AI ​​management function is A registration step for registering a personalized AI or human account managed by another AI management function using the AI ​​management function, A messenger function step that sends and receives conversation requests via the ID or handle name of the aforementioned account, A messenger conversation selection step in which the user is given the option to either have the conversation themselves or have the personalized AI have the conversation in response to the aforementioned conversation request, The system further comprises a conversation linking step that links the conversation between the user or the personalized AI based on the messenger conversation selection step via the ID or handle name of the conversation partner's account. A method for facilitating communication according to any one of claims 1, 2, 3, and 5.

7. A method for facilitating communication using an AI management function, which is performed by a computer, A registration step for registering a personalized AI or human account managed by another AI management function using the AI ​​management function, A messenger function step that sends and receives conversation requests via the ID or handle name of the aforementioned account, A messenger conversation selection step in which the user is given the option to either have the conversation themselves or have the personalized AI have the conversation in response to the aforementioned conversation request, When the personalized AI engages in conversation, the selection acceptance step accepts the selection of at least one of one or more pieces of personalized information stored in the memory unit, wherein the personalized information is at least one of the following: information, models, parameters, settings, and rules for reproducing personality; information, models, parameters, settings, and rules for reproducing characteristics; information, models, parameters, settings, and rules for reproducing preferences; and information, models, parameters, settings, and rules for reproducing voice, or a combination thereof. An application step is to apply the personalized information selected in the selection acceptance step to the AI ​​and operate the personalized AI, A conversation control step in which the personalized AI applied in the above application step is made to perform an automated conversation with a conversation partner as a proxy for the user or as a character managed by the user, A conversation linking step that links the conversation between the user or the personalized AI based on the messenger conversation selection step via the ID or handle name of the conversation partner, A communication-promoting method that incorporates these features.

8. The AI ​​management function is It has a group chat function for personalized AI or humans managed by other AI management functions that use AI management functions, In a group chat, the user can choose whether to converse themselves or have the personalized AI converse. A method for facilitating communication according to any one of claims 1, 2, 3, 5, and 7.

9. A communication facilitation method performed by a computer configured to communicate with a computer or system that performs the communication facilitation method described in Claim 2, An AI management function integration step that links with the personalized AI through the AI ​​management function, An information registration step in which the participating user registers the information of the personalized AI, profile information, matching and automated conversation information, A matching step that matches conversational combinations of the personalized AIs, or the personalized AIs with third-party participants, based on the information of the personalized AI registered in the information registration step, the profile information, and the information for matching and automated conversation. A connection step that connects to the matched AI management function and coordinates the conversation of the personalized AI, A conversation management step that initiates and manages an automated conversation between the matched personalized AIs, or between the personalized AI and a third-party participant, A communication-promoting method that incorporates these features.

10. A communication facilitation method performed by a computer configured to communicate with a computer or system that performs the communication facilitation method described in Claim 3, An AI management function integration step that links with the personalized AI through the AI ​​management function, An information registration step in which the participating user registers the information of the personalized AI, profile information, matching and automated conversation information, A matching step that matches conversational combinations of the personalized AIs, or the personalized AIs with third-party participants, based on the information of the personalized AI registered in the information registration step, the profile information, and the information for matching and automated conversation. A connection step that connects to the matched AI management function and coordinates the conversation of the personalized AI, A conversation management step that initiates and manages an automated conversation between the matched personalized AIs, or between the personalized AI and a third-party participant, A communication-promoting method that incorporates these features.

11. The communication facilitation method, when executing multiple automated conversations, The system includes a multiple conversation presentation step that displays a list of multiple automated conversations currently in progress to the user and presents logs of the content of the multiple automated conversations. This provides a conversation zapping interface that allows users to select a specific conversation session in response to their actions. A method for facilitating communication according to any one of claims 1, 2, 3, 5, 7, 9, and 10.

12. A pause step that, while the personalized AI is performing a conversation with the conversation partner, allows the user to pause the automated conversation by the personalized AI at any time in response to the user's instructions in order to take over the conversation with the conversation partner, A monitoring step in which the personalized AI monitors the conversation between the user and the conversation partner after the pause, The system further includes a restart step in which the paused automated conversation by the personalized AI accepts a restart at any time in response to the user's instructions, and in that case, the conversation is resumed in accordance with the monitored conversation content. A method for facilitating communication according to any one of claims 1, 2, 3, 5, 7, 9, and 10.

13. An advice request processing step in which, during the execution of a conversation between the user and the conversation partner, the user requests advice from the personalized AI regarding the conversation, and the advice request is relayed to the personalized AI, An advice acquisition step involves having the personalized AI analyze the conversation content and the request for advice, and obtaining advice as a result. The system further comprises an advice output step for outputting the aforementioned advice to the user. A method for facilitating communication according to any one of claims 1, 2, 3, 5, 7, 9, and 10.

14. The advice request processing step can receive an advice request via conversation or via an advice button that records the details of what aspects of advice the user wants. The method for facilitating communication according to claim 13.

15. A request / command acquisition step in which a request or command from the user is obtained during the execution of the automated conversation between the personalized AI and the conversation partner, The system further comprises a request-command linking step, which links the request or command acquired in the request-command acquisition step to the personalized AI. A method for facilitating communication according to any one of claims 1, 2, 3, 5, 7, 9, and 10.

16. The advice request processing step is such that, if the conversation is a voice conversation, its content can be obtained by text input from the user, and if the conversation is a text conversation, its content can be obtained by voice input from the user. The advice output step allows the advice to be output as text to the user if the conversation is a voice conversation, and as voice to the user if the conversation is a text conversation. The method for facilitating communication according to claim 13.

17. The request command acquisition step is capable of obtaining text input of the request or command from the user when the automated conversation is a voice conversation, and is capable of obtaining voice input of the request or command from the user when the automated conversation is a text conversation. The method for facilitating communication according to claim 15.

18. The external application request receiving step of the AI ​​management function confirms login or authentication from the external application to the AI ​​management function, and then allows the personalized AI to link to the external application. The method for facilitating communication according to claim 2 or 3.

19. The AI ​​management function includes a character recognition conversation input step of acquiring the conversation content from the text of the conversation partner displayed on the screen of an existing chat service or messenger using a character recognition program and inputting it as a conversation to the personalized AI, The system further includes a conversation output step, which involves integrating the personalized AI with the chat service or messenger to input an automated conversation based on the text it has recognized, thereby executing the automated conversation. A method for facilitating communication according to any one of claims 1, 2, 3, 5, and 7.

20. A calendar registration step that registers in response calendar information a choice between the user themselves or the personalized AI responding, depending on at least one of the date, time, day of the week, another AI management function using the AI ​​management function that will be the conversation partner, an external application, or the handle name or ID of an account, and if the personalized AI responds, which personalized information the personalized AI to which will respond, as well as a setting for a conversation method or message information corresponding to the conversation partner. The calendar automatic response step further comprises: when a conversation request is received from the conversation partner, the system refers to the response calendar information, and if it is determined that the user will respond, it calls the user to initiate the conversation; if it is determined that the personalized AI will respond, the personalized AI will automatically initiate the conversation and have the conversation partner perform the automatic conversation according to the conversation method of the response calendar information set in the calendar registration step, and if message information exists, it sends the message information. A method for facilitating communication according to any one of claims 1, 2, 3, 5, and 7.

21. If the profile information, matching and automated conversation information of all participants in an individual chat being conducted is made public to third-party participants, the privacy setting step includes allowing the chat conversation to be made public to the third-party participants, If the profile information, matching and automated conversation information of all participants in an ongoing chat are available to the third-party participant, the participant control step includes accepting access from the third-party participant, allowing the third-party participant to join the chat and conduct a conversation, It also has, The method for facilitating communication according to claim 9 or 10.

22. A profile matching setting information registration step that allows registration of profile matching setting information, which includes at least one of the following: user profile information (including attribute information or AI growth level), setting information for automated conversation (including conversation format, number of simultaneous conversations, setting to open conversations to third parties, permission to converse with a personalized AI, permission to converse with a human, block conditions or matching conditions), and message information for a specific conversation partner, to the AI ​​management function or an AI-participable chat service, When linking the aforementioned profile and matching settings information from the AI ​​management function to the AI-enabled chat service, the process includes a profile and matching settings information linking step in which the AI ​​management function links the aforementioned profile and matching settings information to the personalized AI-enabled chat service, The aforementioned personalized AI-enabled chat service automatically matches conversational pairs of personalized AIs or between personalized AIs and humans based on the profile and matching settings information. The system further includes a message processing step in which, when the person the conversation partner is identified, a message is sent to the person the conversation partner is based on the message information, or a message is received from the person the conversation partner is and stored. A method for facilitating communication according to any one of claims 1, 2, 3, 5, 7, 9, and 10.

23. A virtual space software or computer game configured to communicate with a computer or system that performs the method described in Claim 2, comprising an AI management function connection step that cooperates with the personalized AI through the AI ​​management function, A communication enhancement method comprising a character conversation coordination step in which, while the user is operating the character of the virtual space software or computer game, conversation can be entrusted to a coordinated, personalized AI.

24. A virtual space software or computer game configured to communicate with a computer or system that performs the method described in Claim 3, comprising an AI management function connection step that cooperates with the personalized AI through the AI ​​management function, A communication enhancement method comprising a character conversation coordination step in which, while the user is operating the character of the virtual space software or computer game, conversation can be entrusted to a coordinated, personalized AI.

25. A virtual space software or computer game configured to communicate with a computer or system that performs the method described in Claim 2, comprising an AI management function connection step that cooperates with the personalized AI through the AI ​​management function, A communication enhancement method that includes an NPC (Non-Player Character) conversation linkage step, which lends personalized AI conversations to NPCs in virtual space software and computer games.

26. A virtual space software or computer game configured to communicate with a computer or system that performs the method described in Claim 3, comprising an AI management function connection step that cooperates with the personalized AI through the AI ​​management function, A communication enhancement method that includes an NPC (Non-Player Character) conversation linkage step, which lends personalized AI conversations to NPCs in virtual space software and computer games.

27. ​​A virtual space software or computer game configured to communicate with a computer or system that performs the method described in Claim 2, comprising an AI management function connection step that cooperates with the personalized AI through the AI ​​management function, A communication facilitation method comprising a post-logout conversation step that allows the personalized AI to perform conversations with a character controlled by the user after the user has finished their operation on the character until the user starts operating on the character again.

28. A virtual space software or computer game configured to communicate with a computer or system that performs the method described in Claim 3, comprising an AI management function connection step that cooperates with the personalized AI through the AI ​​management function, A communication facilitation method comprising a post-logout conversation step that allows the personalized AI to perform conversations with a character controlled by the user after the user has finished their operation on the character until the user starts operating on the character again.

29. After the user has finished operating on the character being operated by the user, an automatic conversation setting step is performed to record the speech pattern, style, content, and how the personalized AI should converse when the character is spoken to by another character, specifying at least one of the date, time, and day of the week, and to have the conversation performed according to the record. The system further includes an automatic action setting step that records where and how the character should act automatically, by specifying at least one of the date, time, and day of the week, and then automatically acts according to that record. The method for facilitating communication according to claim 27 or 28.

30. The post-logout conversation step further includes a post-logout message storage step that allows setting a message for each conversation partner's handle name or ID, A post-logout message delivery step in which the personalized AI transmits a message set for the conversation partner's handle name or ID to the conversation partner, Furthermore, there is a post-logout message receiving step that allows users to receive messages for each handle name or ID, The system further includes a logout message confirmation step that allows the message to be conveyed to the user. The method for facilitating communication according to claim 27 or 28.

31. The AI ​​management function linking step or the AI ​​management function connection step saves login or authentication information to the AI ​​management function, and after logging in or authenticating to the AI ​​management function, it is possible to link an automated conversation from the personalized AI. A method for facilitating communication according to any one of claims 9, 10, 23, 24, 25, 26, 27, and 28.

32. A history storage step of storing the history of conversations between the user and the conversation partner, the content of conversations obtained through monitoring, action information such as requests for advice from the user, or the history of automated conversations by the personalized AI and requests or commands from the user in the storage area of ​​the computer, A learning step in which the saved history or content is used as learning material for the personalized information applied to the personalized AI. A method for facilitating communication according to any one of claims 1, 2, 3, 5, 7, 9, 10, 23, 24, 25, 26, 27, and 28, further comprising:

33. A program for causing a computer to perform the communication facilitation method described in any one of claims 1, 2, 3, 5, 7, 9, 10, 23, 24, 25, 26, 27, and 28.

34. A storage unit for storing one or more pieces of personal information, wherein the personal information is at least one of the following: information, models, parameters, setting information, and rules for reproducing personality; information, models, parameters, setting information, and rules for reproducing characteristics; information, models, parameters, setting information, and rules for reproducing preferences; and information, models, parameters, setting information, and rules for reproducing voice, or a combination thereof, the storage unit. A selection receiving unit that accepts the selection of any of the aforementioned personal information, An application unit that applies the selected personal information to the AI ​​and operates the personalized AI, The conversation control unit, which causes the personalized AI to perform an automated conversation with the conversation partner, An input / output unit that performs automated conversations with the aforementioned conversation partner in voice and text format, A computer equipped with [a specific feature / equipment].

35. A computer-based personalized AI-participatory chat service provider, configured to communicate with an AI management device that manages a personalized AI, A collaboration unit that interacts with the personalized AI through the AI ​​management device, An information registration unit that accepts registration of the personalized AI information, profile information, and information for matching and automated conversation from participating users, A matching unit automatically matches conversational combinations of personalized AIs, or between a personalized AI and a third-party participant, based on the personalized AI information, profile information, and matching and automated conversation information registered in the information registration unit. A connection unit that connects to the matched AI management device and prepares the conversation for the personalized AI, A conversation management unit initiates an automated conversation between the matched personalized AIs, or between the personalized AI and a third-party participant, and manages the conversation. An AI-enabled chat service provider device equipped with [specific features / features].

36. An information processing device that provides a virtual space or computer game composed of a computer, configured to communicate with an AI management device that manages a personalized AI, A collaboration unit that interacts with the personalized AI through the AI ​​management device, In the aforementioned virtual space or computer game, a character conversation coordination unit delegates the conversation processing of the character to a linked, personalized AI while the user is controlling their character, An information processing device equipped with the following features.

Images