Content delivery system, content delivery method, and program

Abstract

Provides appropriate content for a service that allows multiple users to send items of monetary value. [Solution] The input information acquisition unit (101) of the content provision system (1) acquires input information to be input into a generation model (M) capable of generating content (C) provided in a service in which monetaryly valuable items are sent between multiple users. The content generation unit (102) causes the generation model (M) to generate content (C) based on the input information. The content provision unit (103) provides the content (C) generated by the content generation unit (102) to at least one of the multiple users.

Description

Technical Field

[0001] The present disclosure relates to a content providing system, a content providing method, and a program.

Background Art

[0002] Conventionally, services for sending something of monetary value among multiple users have been known. For example, Patent Document 1 describes electronic money as an example of something of monetary value. In the technology of Patent Document 1, a receiving terminal generates receiving information including the amount of electronic money received from another communication terminal, transmits the receiving information to another communication terminal, and when the receiving information is transmitted, receives the transmitted electronic money from another communication terminal, thereby realizing a service for sending electronic money among multiple users.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the technology of Patent Document 1 can only provide a service for sending electronic money among multiple users, and cannot provide appropriate content corresponding to the service to users. This is the same for other things of monetary value other than electronic money. Conventional technologies have not been able to provide appropriate content corresponding to the service for sending something of monetary value among multiple users to users.

[0005] One object of the present disclosure is to provide appropriate content corresponding to a service for sending something of monetary value among multiple users.

Means for Solving the Problems

[0006] The content provision system relating to this disclosure includes: an input information acquisition unit that acquires input information to be input into a generation model capable of generating content provided in a service in which monetary value is sent between multiple users; a content generation unit that causes the generation model to generate the content based on the input information; and a content provision unit that provides the content generated by the content generation unit to at least one of the multiple users. [Effects of the Invention]

[0007] This disclosure enables the provision of appropriate content for services that facilitate the exchange of monetaryly valuable items between multiple users. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows an example of the hardware configuration of a content delivery system. [Figure 2] This figure shows an example of a screen displayed on the user terminal of the sending user. [Figure 3] This figure shows an example of a screen displayed on the recipient user's terminal. [Figure 4] This figure shows an example of the functions implemented in the content delivery system. [Figure 5] This figure shows an example of a user database. [Figure 6] This figure shows an example of input and output for a generative model. [Figure 7] This figure shows an example of the process performed by the content delivery system. [Figure 8] This figure shows an example of a function that can be achieved through modification. [Figure 9] This figure shows an example of the input and output of the generative model in Modification Example 1. [Figure 10] This figure shows an example of the input and output of the generative model in Modification Example 2. [Figure 11]This figure shows an example of the input and output of the generative model in Modification Example 3. [Figure 12] This figure shows an example of the input screen in modified example 4. [Figure 13] This figure shows an example of the input and output of the generative model in Modification Example 5. [Figure 14] This figure shows an example of the input and output of the generative model in Modification Example 7. [Figure 15] This figure shows an example of the input and output of the generative model in modified example 8. [Modes for carrying out the invention]

[0009] [1. Hardware configuration of the content delivery system] This document describes an example of an embodiment of the content provision system, content provision method, and program related to this disclosure. Figure 1 shows an example of the hardware configuration of the content provision system. For example, the content provision system 1 includes a server 10 and a user terminal 20. Each of the server 10 and the user terminal 20 is connected to a network N such as the Internet or a LAN. At least one of the server 10 and the user terminal 20 may be present in multiple units.

[0010] Server 10 is a server computer. For example, Server 10 includes a control unit 11, a storage unit 12, and a communication unit 13. The control unit 11 includes at least one processor. The storage unit 12 includes at least one of volatile memory such as RAM and non-volatile memory such as flash memory. The communication unit 13 includes at least one of a communication interface for wired communication and a communication interface for wireless communication.

[0011] The user terminal 20 is the user's computer. For example, the user terminal 20 is a smartphone, a tablet, a personal computer, or a wearable terminal. The user terminal 20 includes a control unit 21, a storage unit 22, a communication unit 23, an operation unit 24, and a display unit 25. The hardware configurations of the control unit 21, the storage unit 22, and the communication unit 23 may be the same as those of the control unit 11, the storage unit 12, and the communication unit 13, respectively. The operation unit 24 is an input device such as a touch panel or a mouse. The display unit 25 is a display such as a liquid crystal or an organic EL.

[0012] Note that the programs stored in the storage units 12 and 22 may be supplied to the server 10 or the user terminal 20 via the network N. Further, at least one of a reading unit (e.g., a memory card slot) for reading a computer-readable information storage medium and an input / output unit (e.g., a USB port) for inputting / outputting data to / from an external device may be included in the server 10 or the user terminal 20. For example, a program stored in the information storage medium may be supplied to the server 10 or the user terminal 20 via at least one of the reading unit and the input / output unit.

[0013] Further, the content providing system 1 may include at least one computer. The computer included in the content providing system 1 is not limited to the example of FIG. 1. For example, the content providing system 1 may include only the server 10. In this case, the user terminal 20 exists outside the content providing system 1. The content providing system 1 may include the server 10 and another computer not shown in FIG. 1.

[0014] In this embodiment, the content providing system 1 can communicate with an external system 2 via a network N. The external system 2 includes at least one server computer. The hardware configuration of the server computer of the external system 2 may be the same as that of the server 10. In this embodiment, a case where the entity that manages the server 10 (for example, an operator who operates the settlement service described later) and the entity that manages the server computer of the external system 2 (for example, an operator who operates the service of the generation model described later) are different will be described, but these entities may be the same.

[0015] [2. Overview of the Content Providing System] In this embodiment, a case where the content providing system 1 is applied to a settlement service will be exemplified. The settlement service is a service that provides electronic settlement (cashless settlement) to each of a plurality of users. A user can use any settlement means in the settlement service. For example, a user may use electronic money, a balance not called electronic money, a credit card, a card other than a credit card, points, a bank account, an account other than a bank account, a cryptocurrency, a wallet, or other settlement means in the settlement service. Since the settlement means may also be used for payment, it can also be called a payment means.

[0016] In this embodiment, a case where a user uses the settlement service from a settlement app stored in the user terminal 20 will be exemplified. The settlement app is an application provided by the administrator of the settlement service. A user may use the settlement service from an app other than the settlement app stored in the user terminal 20 (for example, an app provided by a credit card company or other non-settlement apps), or may use the settlement service from a browser stored in the user terminal 20. A user may use the settlement service with the IC chip of the user terminal 20 instead of a program stored in the user terminal 20.

[0017] For example, a payment service provides users with multiple functions. In this embodiment, we will take the money transfer function, which allows a user to send electronic money to another user, as an example of one of the multiple functions of a payment service. Hereafter, the user who sends electronic money to another user will be called the sending user. The other user who receives electronic money from the sending user will be called the receiving user. In Figure 1, only one user terminal 20 is shown, but in reality, there will be a user terminal 20 for the sending user and a user terminal 20 for the receiving user. When the sending user and the receiving user are not distinguished, they will simply be referred to as users.

[0018] Figure 2 shows an example of a screen displayed on the sender user's terminal 20. For example, when the sender user operates their own terminal 20 to launch the payment app, the sender user's terminal 20 displays the top screen SC1 of the payment app on the display unit 25, as shown in the upper left of Figure 2. When the sender user selects button B10 to use the money transfer function, the sender user's terminal 20 displays the input screen SC2, which accepts input of necessary information for the money transfer function, on the display unit 25, as shown in the upper right of Figure 2. In the example in the upper right of Figure 2, a list of other users who are candidates for the recipient user is displayed on the input screen SC2.

[0019] For example, other users who could be candidates for the recipient may be other users registered in the contact list stored on the sender user's user terminal 20, or other users registered with the payment service (for example, recipient users to whom the sender user has previously sent electronic money). The sender user may also specify other users not displayed on the input screen SC2 as recipient users. The sender user may also specify a recipient user by searching for the recipient user's account ID or phone number, etc. The sender user may also send a receiving link to the recipient user using a means other than the payment app (for example, email or a messaging app) without specifically specifying a recipient user on the input screen SC2.

[0020] For example, when the sending user specifies the recipient user on the input screen SC2 in the upper right of Figure 2, the sending user's user terminal 20 displays the input screen SC2 on the display unit 25 for receiving input details of the electronic money to be sent to the recipient user, as shown in the lower left of Figure 2. In the example in the lower left of Figure 2, the input screen SC2 includes an input form F20 for receiving an amount and an input form F21 for receiving a message to the recipient user. The sending user enters an arbitrary amount in input form F20. The sending user enters an arbitrary message in input form F21. The sending user may omit entering a message in input form F21.

[0021] In this embodiment, content C is generated based on a generation model managed by an external system 2, which is displayed at least when the sending user sends electronic money to the receiving user, and when the receiving user receives electronic money from the sending user. In this embodiment, the case where content C is displayed at both of these times is given as an example, but content C may be displayed at only one of these times. Details of the generation model will be described later.

[0022] In the lower left example of Figure 2, the input screen SC2 includes an input form F22 that accepts input for the content C desired by the sending user. For example, the sending user enters text such as "I would like an image with lots of flowers" into input form F22. The sending user may omit inputting anything into input form F22. In this case, the generation model does not have to generate content C, or it may generate background information based on other information, as shown in the modified example described later.

[0023] For example, when a sending user enters the desired content C into input form F22 and selects button B23, the generation model generates content C based on the content entered in input form F22. Once the generation model generates content C, as shown in the lower right of Figure 2, the sending user's terminal 20 displays a confirmation screen SC3 on the display unit 25 for the sending user to make a final confirmation. The confirmation screen SC3 includes content C. When the sending user slides the slider bar B30, electronic money is sent from the sending user to the receiving user. The receiving user launches the payment application stored on their user terminal 20 and performs the operation to receive the electronic money.

[0024] Figure 3 shows an example of a screen displayed on the recipient user's terminal 20. For example, when the recipient user operates their own terminal 20 and launches the payment application, the recipient user's terminal 20 displays a receiving screen SC4 on the display unit 25 for the recipient user to receive the electronic money, as shown in Figure 3. The receiving screen SC4 includes content C generated by the generation model. When the recipient user selects button B40, the receipt of the electronic money is completed. The receiving screen SC4 may also accept a message from the recipient user to the sender user. Based on this message, content C may be generated by the generation model.

[0025] As described above, the content provision system 1 of this embodiment displays the content C generated by the generation model on both the confirmation screen SC3 displayed on the user terminal 20 of the sending user and the receiving screen SC4 displayed on the user terminal 20 of the receiving user. This enables the content provision system 1 to provide appropriate content C according to the payment service for sending electronic money between multiple users. The details of the content provision system 1 will be described below.

[0026] [3. Functions implemented in the content delivery system] Figure 4 shows an example of the functions implemented in the content provision system 1. The various components implemented in the content provision system 1 of this embodiment can be configured by combining them into a single device or by further distributing them among multiple devices.

[0027] [3-1. Functions implemented by the server] For example, server 10 includes a data storage unit 100, an input information acquisition unit 101, a content generation unit 102, and a content provision unit 103. The data storage unit 100 is implemented by a storage unit 12. Each of the input information acquisition unit 101, the content generation unit 102, and the content provision unit 103 is implemented by a control unit 11.

[0028] [Data Storage Unit] The data storage unit 100 stores various types of data related to the payment service. For example, the data storage unit 100 stores the user database DB1.

[0029] Figure 5 shows an example of a user database DB1. User database DB1 is a database that stores various information for multiple users. For example, user database DB1 stores user ID, login account, password, basic user information, payment method information, and usage history information. The information stored in user database DB1 is not limited to the example in Figure 5. User database DB1 may store other information. For example, user database DB1 may store the history of content C that has been generated in the past, and information on users who are candidates for recipient users.

[0030] A user ID is an example of user identification information that can identify a user. A login account is also an example of user identification information. A login account is user identification information that a user enters to log in to a payment service. The login account may be freely changed by the user. A user ID is user identification information that is managed separately from the login account. In this embodiment, we take the example of a case where there is a login account separate from the user ID (where there are at least two pieces of user identification information), but there may be only one piece of user identification information. That is, the user ID and login account do not have to be separate. A password is authentication information that is verified during login. The login account and password may be the user's biometric information.

[0031] User basic information is fundamental information about a user. Users may be able to edit their own user basic information. User basic information may also be known as demographic information. User basic information may indicate the user's attributes. For example, user basic information may include the user's name, gender, date of birth, age or age group, address, preferences, occupation, place of employment, type of services used, or a combination of these.

[0032] Payment method information is information that identifies the payment methods available to a user in the payment service. For example, payment method information may include credit card numbers, electronic money numbers, bank account information, or point card numbers. Payment method information may also indicate the balance of electronic money. When a sending user sends electronic money to a receiving user, the sending user's electronic money balance decreases. When a receiving user receives electronic money from a sending user, the receiving user's electronic money balance increases. The current balance of electronic money may be managed on a computer other than server 10.

[0033] Usage history information is information about a user's usage history of the payment service. For example, usage history information may show the name of the store where the user made a payment using the payment service, the payment amount, and the date and time of the payment. Each user's usage history information may show the amount of electronic money that the user sent to a recipient user as the sender, the recipient user's information, and the date and time of the transfer. Each user's usage history information may also show the amount of electronic money that the user received from a sender user as the recipient, the sender user's information, and the date and time of receipt. Each user's usage history information is updated each time the user uses the payment service.

[0034] The data stored in the data storage unit 100 is not limited to the above example. The data storage unit 100 may store other data related to the payment service. For example, the data storage unit 100 may store data necessary for displaying the top screen SC1, input screen SC2, confirmation screen SC3, receipt screen SC4, or other screens. In this embodiment, the case in which the external system 2 stores the actual data of the generated model is given as an example, but the data storage unit 100 may also store the actual data of the generated model. In this case, the external system 2 does not need to exist.

[0035] The generative model is a model capable of generating content C provided in a service where items of monetary value are sent between multiple users. In this embodiment, the example given is that the "item" in the item of monetary value is an intangible object, but the "item" may also be a tangible object. If the item of monetary value is an intangible object, the item of monetary value may also be electronic data. In this case, the item of monetary value can also be called electronic value. The item of monetary value may be an asset with asset value, or it may not be classified as an asset in particular.

[0036] In this embodiment, electronic money is described as an example of something with monetary value. Therefore, wherever "electronic money" is written, it can be read as "something with monetary value." Something with monetary value can be anything with some kind of value, and is not limited to electronic money. For example, something with monetary value may be a balance not called electronic money, cryptocurrency, coupons, points, electronic tickets (e.g., tickets for admission to sports or events), digital tokens, electronic gifts, product vouchers, service vouchers, electronic content, or other things. Something with monetary value may be the item itself (e.g., a product) or the service itself (e.g., a service provided at a store).

[0037] A service that allows multiple users to send something of monetary value is a service that provides a function for sending something of monetary value to each of the multiple users. In this embodiment, a payment service is described as an example of such a service. The service may be any service that has a function for sending something of monetary value, and may be other services. For example, the service may be a financial service, an e-commerce service, an SNS (Social Networking Service), a messaging app service, a chat service, an SMS (Short Message Service), a ticket sales service, a communication service, a travel booking service, or other services. Where it says "payment service," it can be read as any of these services.

[0038] In this embodiment, the generation model includes a program that shows a series of processes for generating content C based on information input to it, and parameters referenced by the program. Hereafter, the information input to the generation model will be referred to as input information. The generation model may not include any parameters, and the program itself may constitute the generation model. Alternatively, the parameters may be incorporated as part of the program of the generation model. The generation model may also include other information besides the program and parameters.

[0039] In this embodiment, we take the example of a generative AI (Artificial Intelligence) capable of generating content C corresponding to a generative model. The generative AI is an AI capable of generating electronic products based on input information input to itself. There are various definitions of generative AI, but the generative AI in this embodiment may be a generative AI in various known definitions. For example, the generative AI may be an AI developed not only by machine learning methods, but also by other methods other than machine learning. The generative AI may be a single-modal type AI capable of processing only a specific format of input information, or a multimodal type AI capable of processing multiple formats of input information.

[0040] Content C is information (data) provided electronically to the user. Content C is not limited to visual information; it may also be auditory information. In this embodiment, as an example of Content C, a background image corresponding to the background on the screen of a payment application is described. Content C can be any information provided to the user, and is not limited to a background image. For example, Content C may be other images besides a background image (e.g., an image corresponding to the foreground), video, audio, email, other messages besides email (e.g., messages from a messaging application), notifications in an application such as a payment application, or other information.

[0041] For example, the generative model may be a large-scale language model, a machine learning model that does not fall under the category of a large-scale language model (a model that has been trained using machine learning techniques, such as a GAN (Generative Adversarial Network) or a VAE (Variational Autoencoder)), or another model. For example, the generative model may be a Dall-E series model, MidJourney, Stable Diffusion, Artbreeder, Deep Dream Generator, or a generative AI (so-called image generation AI) such as Runway ML.

[0042] Furthermore, the generative model may include a large-scale language model (for example, a transformer-type model such as GPT (Generative Pre-trained Transformer), or other models such as neural networks that are not classified as transformers) and other models that generate content C in cooperation with the large-scale language model (for example, models in the Dall-E series). In other words, the generative model may consist of multiple models. Also, the generative model may not be called a generative AI. For example, a machine learning model that is not called a generative AI may correspond to the generative model.

[0043] Furthermore, the generative model is assumed to have been trained with various content C as training data. If the generative model is a model capable of natural language processing, such as a large-scale language model, then the generative model is assumed to have been trained with various texts as training data. The parameters of the generative model are adjusted by training with the training data. A pre-trained general-purpose generative model may be used as is, or a generative model fine-tuned for the content delivery system 1 may be used.

[0044] Furthermore, the parameters of the generative model may be the same as those of known parameters. For example, the parameters of the generative model may be weights and biases. The generative model may include multiple layers, such as an input layer, a hidden layer, and an output layer. The method for learning the generative model may be the same as that of known machine learning techniques. Similar to the definition of AI, there are various definitions of machine learning, but the machine learning in this embodiment includes various known definitions. For example, deep learning is also included in machine learning.

[0045] [Input Information Acquisition Unit] Figure 6 shows an example of the input and output of a generative model. Hereafter, each function of the input information acquisition unit 101, etc., will be explained with reference to Figure 6. The input information acquisition unit 101 acquires the input information that is input to the generative model. In this embodiment, an example is given where the input information indicates text written in natural language, but the input information may be in any format that the generative model can process. For example, the input information may be natural language text that is not in sentence form, non-natural language text, images, videos, documents or other files, other formats, or a combination thereof. Note that the text is not limited to characters, may be numbers, or other symbols other than characters and numbers.

[0046] The input information acquisition unit 101 can acquire input information from any location. For example, if the input information is entered by a user, the input information acquisition unit 101 acquires the input information from the user terminal 20. If the input information is already stored in the data storage unit 100, the input information acquisition unit 101 acquires the input information from the data storage unit 100. If the input information is stored in an information storage medium, the input information acquisition unit 101 acquires the input information from the information storage medium. If the input information is stored in a computer other than the server 10 and the user terminal 20, the input information acquisition unit 101 acquires the input information from the other computer. The input information acquisition unit 101 may combine the information acquired from each of these multiple locations and acquire it as input information.

[0047] In this embodiment, we take the example that the generation model is a model of a generation AI capable of generating content C. Therefore, the input information acquisition unit 101 acquires a prompt as input information for instructing the generation model to generate content C. The prompt is an instruction to the generation model. The prompt may be text written in a natural language that humans can understand, or it may be other information other than such text. For example, the prompt may be text written in a language other than natural language (e.g., program code written in a programming language, or structured text written in a markup language), an image, a file such as a document, or other information. The prompt may be a combination of these pieces of information.

[0048] In this embodiment, the input information acquisition unit 101 acquires user input information related to input received on the input screen SC2 for sending or requesting electronic money by any of the multiple users, and acquires a prompt based on the user input information. User input information is the input information entered by the user. In the upper right and lower left examples of Figure 2, the input screen SC2 for a sending user to send electronic money to a receiving user is shown, so the user input information is the input information entered by the sending user. The user input information may show all the information entered on the input screen SC2, or it may show some of the information entered on the input screen SC2.

[0049] For example, the input information acquisition unit 101 may acquire user input information indicating the content of content C entered by any of the multiple users on the input screen SC2. The content of content C can also be the details or characteristics of content C. For example, the content of content C may be the object, color, pattern, brightness, image shown in content C, or a combination thereof. On the input screen SC2, an image showing the general color scheme may be displayed, and the user input information may indicate the image selected from among them.

[0050] In this embodiment, we take the example of a case where the user input information indicates the content of Content C desired by the sending user, but the user input information may also indicate the content of Content C desired by the receiving user. That is, the content of Content C may be entered by the sending user or by the receiving user.

[0051] In the lower left example of Figure 2, the sending user enters the content of Content C into input form F22. Therefore, the user input information represents the text entered into input form F22. If the sending user enters text such as "I would like an image with lots of flowers" into input form F22, the user terminal 20 sends user input information representing that text to the server 10. The input information acquisition unit 101 acquires the user input information from the user terminal 20.

[0052] In this embodiment, we take the example of a case where the user input information only indicates the content of Content C, but the user input information may indicate the content of Content C and other information other than the content of Content C. The user input information may indicate only other information without indicating the content of Content C. An example of other information will be explained in the modified examples below. The user input information may indicate multiple pieces of other information. The user input information may indicate the content of Content C and multiple pieces of other information.

[0053] In this embodiment, the input information acquisition unit 101 acquires pre-prepared input information and acquires a prompt based on the pre-prepared input information. The pre-prepared input information is pre-prepared input information. The pre-prepared input information is not user input information entered by the user, but input information prepared by the content provision system 1. When the pre-prepared input information is used as all or part of the prompt, it can also be called the default prompt. The pre-prepared input information is at least part of the prompt.

[0054] Furthermore, if the model is not specialized in generating content C, but is a general-purpose model capable of generating other products (e.g., text) besides content C, then the content provision system 1 needs to instruct the generation model to generate content C. For this reason, the prepared input information may indicate the content of the processing that the generation model should perform. The processing that the generation model should perform can also be called the generation model's task. In this embodiment, the generation of content C corresponds to the task. For example, the prepared input information may indicate text written in natural language that the generation model should generate content C based on a prompt.

[0055] For example, the input information acquisition unit 101 may acquire prepared input information indicating that the generation model will generate content C according to the user input information. The prepared input information consists of text instructing the generation model to generate content C according to the user input information. In the example in Figure 6, the prepared input information consists of text such as, "You are the AI ​​that generates content when sending electronic money. Please generate appropriate content based on the following user input information." The prepared input information may also indicate that a background image should be generated as content C.

[0056] Note that "[User Input Information]" in Figure 6 is a tag indicating the location where user input information is set. Therefore, the input information acquisition unit 101 obtains the final prompt by setting the user input information in the portion of the prepared input information indicated by the tag. Setting user input information means inserting user input information. User input information may be input into the generation model separately from the prepared input information, rather than being incorporated into the prepared input information.

[0057] The input information acquisition unit 101 may acquire user input information directly as a prompt without acquiring prepared input information. The input information acquisition unit 101 may acquire a prompt without acquiring user input information. For example, the input information acquisition unit 101 may acquire input information, as described in the modified example below, as a prompt without acquiring user input information. If the generative model is not a generative AI, the input information acquisition unit 101 may acquire input information that is not called a prompt.

[0058] Furthermore, instead of the sending user voluntarily sending electronic money to the receiving user, the receiving user may request electronic money from the sending user. In other words, the sending user may send electronic money to the receiving user in response to a request from the receiving user. In this case, the information entered by the receiving user when requesting electronic money from the sending user corresponds to the user input information.

[0059] For example, a screen in which a recipient user enters the necessary information to request electronic money from a sender user corresponds to "an input screen for any of multiple users to request electronic money." On this screen, at least one of the following may be accepted as input: the amount of electronic money requested from the sender user, a message from the recipient user to the sender user, and the content C desired by the recipient user. User input information may be any of this information entered on the screen.

[0060] The input information may be any information input into the generation model, and is not limited to the examples of this embodiment. For example, the input information may include only user input information, or only prepared input information. The input information may include information of the modified examples described later, or it may indicate the relationships between multiple users to whom electronic money is sent. If a sample of content C is stored in the data storage unit 100, the input information may include the sample. For example, a method called RAG (Retrieval Augmented Generation), which utilizes samples stored in a database, may be used to generate content C. The input information may also be information on past remittances.

[0061] Furthermore, if the generation model is specialized in generating content C, it can generate content C even without prompting, so the generation of content C does not need to be indicated by a prompt. Even if the generation model is a general-purpose model not specialized in generating content C, if the generation of content C can be configured in the generation model settings, it can generate content C even without prompting, so the generation of content C does not need to be indicated by a prompt.

[0062] [Content Generation Section] The content generation unit 102 causes the generation model to generate content C based on the input information. The content generation unit 102 causes the generation model to generate content C by inputting the input information to the generation model. In this embodiment, a prompt is acquired as input information, so the content generation unit 102 causes the generation model to generate content C based on the prompt. The content generation unit 102 causes the generation model to generate content C by inputting the prompt to the generation model. The content generation unit 102 acquires the content C generated by the generation model.

[0063] In this embodiment, since the actual data of the generation model is stored in the external system 2, the content generation unit 102 causes the generation model to generate content C by sending input information to the external system 2. When the external system 2 obtains input information from the server 10, it inputs the input information into the generation model it manages. The external system 2 sends the content C generated by the generation model to the server 10. The content generation unit 102 obtains the content C generated by the generation model from the external system 2. If the actual data of the generation model is stored in the data storage unit 100, the content generation unit 102 can cause the generation model to generate content C by inputting input information to the generation model stored in the data storage unit 100.

[0064] The processing performed by the generative model may be the same as known processing. The generative model may generate content C corresponding to the input information by processing similar to known generative AI. For example, the generative model computes an embedding representation of the input information based on trained parameters. The embedding representation is information that the generative model uses to recognize meaning. For example, the embedding representation may be a multidimensional vector, or it may be in a form other than a multidimensional vector (e.g., array form, matrix form, multiple numbers, single number, or feature map). If the input information includes an image, a convolution process may be performed.

[0065] For example, a generative model outputs content C corresponding to an embedding representation of the input information. The generative model may divide the input information into processing units called tokens and then compute the embedding representation of each token. Based on the sequence of embedding representations of the individual tokens, the generative model predicts the continuation as needed and then outputs content C. The generative model may include an input layer that accepts input information, an intermediate layer that computes the embedding representations, and an output layer that outputs content C corresponding to the embedding representations. A series of processes may be performed to generate content C from the input information based on each of these layers. Parameters may be associated with each of these layers.

[0066] In the example in Figure 6, if the sending user inputs user information such as "I would like an image with lots of flowers," the generative model recognizes from the embedded representation of the user input that it is instructed to generate an image showing lots of flowers. The generative model then generates and outputs content C showing lots of flowers as content C corresponding to the embedded representation. The generative model is pre-trained with various flower images for training, and it is assumed that the generative model is capable of generating flower images through this prior training.

[0067] The generative model only needs to generate at least one piece of content C based on one input. The generative model may generate only one piece of content C, or it may generate multiple pieces of content C. The number of pieces of content C that the generative model should generate may be indicated in the input information. For example, the number of pieces of content C that the generative model should generate may be indicated in the user input information or in the prepared input information. The generative model may also generate multiple types of content C, such as a background image and a video. The types of content C that the generative model should generate may be indicated in the input information. For example, the types of content C that the generative model should generate may be indicated in the user input information or in the prepared input information.

[0068] Alternatively, the content generation unit 102 may cause the generation model to generate content C such that the content C provided to the sending user and the content C provided to the receiving user are different, rather than having content C common to both the sending user and the receiving user. In this case, the content generation unit 102 causes the generation model to generate a first content C provided to the sending user. The first content C is not provided to the receiving user. The content generation unit 102 causes the generation model to generate a second content C, which is different from the first content C, and is provided to the receiving user. The second content C is not provided to the sending user.

[0069] Furthermore, the content generation unit 102 may have the generation model modify the content C. For example, the sending user may input the modification details for the content C displayed on the confirmation screen SC3. For example, the sending user can instruct the content C displayed on the confirmation screen SC3 to be modified or changed to another content C. In this case, the modification details correspond to the user input information at the time of modification. The content generation unit 102 inputs the modification details into the generation model, and the generation model modifies the content C. The modification of content C may be performed by a process similar to modification in known generation AI. The content generation unit 102 obtains the modified content C from the generation model. The content generation unit 102 may also have the generation model regenerate a completely new content C instead of modifying the content C.

[0070] [Content Provision Department] The content provision unit 103 provides content C generated by the content generation unit 102 to at least one of the multiple users. At least one of the multiple users is either a sending user or a receiving user. The multiple users may include multiple sending users or multiple receiving users. Providing content C means sending content C to the user terminal 20. In this embodiment, the content provision unit 103 provides content C on a payment application screen as an example, but the content provision unit 103 may also provide content C using email, messaging applications, social networking services (SNS), the notification function of the user terminal 20, or other means.

[0071] In the lower right example of Figure 2, the content provider 103 provides content C, an example of content C, to the sending user by sending display data of confirmation screen SC3, which includes content C, to the sending user's terminal 20. In the example of Figure 3, the content provider 103 provides content C to the receiving user by sending display data of receiving screen SC4, which includes content C, to the receiving user's terminal 20. The content provider 103 may also provide content C on a history screen showing the history of past remittances, a notification function screen in a payment application, or another screen.

[0072] Furthermore, the content C provided to the sending user and the content C provided to the receiving user may be different. In this case, the content provider 103 may provide the first content C to the sending user and a second content C different from the first content C to the receiving user. Alternatively, the content provider 103 may provide content C only to the sending user. Alternatively, the content provider 103 may provide content C only to the receiving user. Alternatively, the content provider 103 may provide content C to other users different from the sending user and the receiving user.

[0073] [3-2. Functions implemented on the user terminal] For example, the user terminal 20 includes a data storage unit 200, an operation reception unit 201, and a display control unit 202. The data storage unit 200 is implemented by a storage unit 22. The operation reception unit 201 and the display control unit 202 are implemented by a control unit 21.

[0074] [Data Storage Unit] The data storage unit 200 stores the data necessary for the user to use the payment service. For example, the data storage unit 200 stores the payment application. If the user uses the payment service from another application or browser instead of the payment application, the data storage unit 200 stores the other application or browser.

[0075] [Operation Reception Section] The operation reception unit 201 receives various user operations. For example, the operation reception unit 201 receives operations on the payment application. The operation reception unit 201 transmits data indicating the user's operation to the server 10.

[0076] [Display Control Unit] The display control unit 202 displays various screens on the display unit 25. For example, the display control unit 202 displays the top screen SC1, input screen SC2, confirmation screen SC3, and receiving screen SC4 on the display unit 25. The display control unit 202 communicates with the server 10 to receive display data for these screens and displays these screens on the display unit 25 based on the display data. The display data is the data necessary for displaying the screen. The display data may be data for the entire screen or data for a part of the screen. For example, the display data may be data in a markup language such as HTML, image data, text data, or other data.

[0077] [4. Processes executed by the content delivery system] Figure 7 shows an example of the processing performed by the content provision system 1. In Figure 7, the processing related to the generation of content C is shown among the processing performed by the content provision system 1. The processing in Figure 7 is performed by the control units 11 and 21 executing the programs stored in the storage units 12 and 22, respectively. The steps in Figure 7 are an example of a content provision method. Figure 7 shows the processing performed when a sending user sends electronic money to a receiving user.

[0078] As shown in Figure 7, when the payment application is launched, the sending user's terminal 20 performs a login process with the server 10 to allow the sending user to log in to the payment service (S1). If the login process is successful, the server 10 performs a process with the sending user's terminal 20 to display the top screen SC1 (S2). When the sending user selects button B10, the server 10 performs a process with the sending user's terminal 20 to display the input screen SC2 (S3).

[0079] In S3, Server 10 retrieves phonebook data from User Terminal 20 as needed, generates display data for input screen SC2, and sends it to User Terminal 20. User Terminal 20 displays input screen SC2 in the upper right of Figure 2. When the sending user selects a recipient user, User Terminal 20 sends the user identification information of the selected recipient user to Server 10. After Server 10 obtains the user identification information of the recipient user from User Terminal 20, it sends the display data for input screen SC2 in the lower left of Figure 2 to User Terminal 20. User Terminal 20 accepts inputs to input form F20, etc.

[0080] When the sending user enters the necessary information on the input screen SC2 in the lower left of Figure 2, the sending user's terminal 20 transmits the user input information to the server 10 (S4). In this embodiment, the content C entered in the input form F22 corresponds to the user input information, so the user terminal 20 transmits the user input information indicating the content C entered by the sending user to the server 10. The user terminal 20 also transmits other information, such as the content entered in the input form F20, to the server 10.

[0081] Server 10 obtains user input information from the source user's user terminal 20 (S5). Server 10 obtains prepared input information stored in the storage unit 12 (S6). Server 10 obtains a prompt as input information based on the user input information and the prepared input information (S7). In S7, Server 10 obtains a prompt by setting the user input information in a predetermined position in the prepared input information. Server 10 sends the prompt as input information to the external system 2, causing the generation model to generate content C (S8). In S8, Server 10 obtains the content C generated by the generation model from the external system 2.

[0082] Server 10 performs a process to display a confirmation screen SC3 containing content C with the user terminal 20 of the sending user (S9). In S9, Server 10 provides content C to the sending user by sending the display data of the confirmation screen SC3 containing content C to the user terminal 20. When the sending user slides the slider bar B30, Server 10 performs a process to send electronic money from the sending user to the receiving user with the user terminal 20 (S10). The process in S10 may be the same as the process used in known payment services. In S10, Server 10 updates the usage history information of both the sending user and the receiving user so that each indicates that an electronic money transfer has taken place. At the time of S10, the sending user's electronic money balance may increase and the receiving user's electronic money balance may decrease.

[0083] When the payment application is launched, the recipient user's terminal 20 performs a login process with the server 10 to allow the recipient user to log in to the payment service (S11). If the login process is successful, the server 10 performs a process with the recipient user's terminal 20 to display the receiving screen SC4 containing content C (S12). The content C generated in S8 is assumed to be stored in the user database DB1. In S12, the server 10 generates display data for the receiving screen SC4 containing content C stored in the user database DB1 and sends it to the recipient user's terminal 20. When the recipient user selects button B40, the server 10 performs a process with the recipient user's terminal 20 to receive the electronic money sent from the sender user (S13), and this process ends. The process in S13 may be the same as the process used in known payment services.

[0084] [5. Summary of Embodiments] The content provision system 1 of this embodiment acquires input information that is input to the generation model. Based on the input information, the content provision system 1 causes the generation model to generate content C. The content provision system 1 provides content C to at least one of the multiple users. As a result, the content provision system 1 can provide appropriate content C according to the payment service by providing content C generated by the generation model, rather than providing images etc. prepared by the administrator. For example, if the confirmation screen SC3 has no background, the sending user may feel that the confirmation screen SC3 is bland and unsatisfying, but since content C generated by the generation model is provided to the sending user, the sending user can feel that the confirmation screen SC3 has a good design. If the receiving screen SC4 has no background, the receiving user may feel that the receiving screen SC4 is bland and unsatisfying, but since content C generated by the generation model is provided to the receiving user, the receiving user can feel that the receiving screen SC4 has a good design. The content provision system 1 can also add value to the payment service by providing content C. Since users can communicate with others through content C, the content delivery system 1 can also enhance user interest.

[0085] Furthermore, the generative model is a model of a generative AI capable of generating content C. The content provision system 1 acquires prompts as input information to instruct the generative model to generate content C. Based on the prompts, the content provision system 1 causes the generative model to generate content C. As a result, the content provision system 1 can control the content C generated by the generative AI using prompts, thereby providing more appropriate content C.

[0086] Furthermore, the content provision system 1 acquires user input information related to inputs received on the input screen SC2 for sending or requesting electronic money by any of the multiple users. Based on the user input information, the content provision system 1 acquires prompts. As a result, the content provision system 1 can provide content C that more appropriately reflects the user's intent by utilizing the user input information directly entered on the input screen SC2.

[0087] Furthermore, the content provision system 1 acquires user input information indicating the content C entered by one of the multiple users on the input screen SC2. This allows the content provision system 1 to provide appropriate content C that directly reflects the content C desired by the user. For example, the content provision system 1 can generate content that aligns with the intentions of the sending user based on the user input information entered by the sending user. The content provision system 1 can also generate content that aligns with the intentions of the receiving user based on the user input information entered by the receiving user.

[0088] Furthermore, the content provision system 1 acquires pre-prepared input information and obtains prompts based on this pre-prepared input information. This allows the content provision system 1 to provide appropriate content C using the pre-prepared input information, even if the user does not input all of the prompt's content. For example, even if the user input information alone is insufficient as a prompt, the content provision system 1 can supplement the information necessary to generate content C using the pre-prepared input information.

[0089] Furthermore, the content provision system 1 acquires prepared input information indicating that the generation model should generate content C according to the user input information. This allows the content provision system 1 to instruct the generation model to generate content C according to the user input information based on the prepared input information, thereby enabling more reliable generation of content C according to the user input information.

[0090] [6. Variant] This disclosure is not limited to the embodiments described above. This disclosure may be modified as appropriate without departing from the spirit of this disclosure.

[0091] Figure 8 shows an example of a function realized by a modified example. For example, the server 10 includes an approval request unit 104 and a learning unit 105. Each of the approval request unit 104 and the learning unit 105 is realized by the control unit 11.

[0092] [6-1. Variation 1] For example, the input information acquisition unit 101 may acquire ready input information indicating that the generation model will generate content C corresponding to past content previously generated by the generation model. Past content is content C previously provided to at least one of multiple users. Content C corresponding to past content is content C that is the same as past content, content C that is not the same as past content but shares characteristics with past content, or content C that is not the same as past content but has different characteristics with past content. Here, characteristics refer to visual characteristics. For example, characteristics are objects, colors, patterns, brightness, images, or combinations thereof shown in past content or content C.

[0093] For example, the user may select from the following: first prepared input information containing wording that instructs the generation model to generate content C which is the same as past content; second prepared input information containing wording that instructs the generation model to generate content C which is not the same as past content but shares common characteristics with past content; and third prepared input information containing wording that instructs the generation model to generate content C which is not the same as past content but has different characteristics with past content. The input information acquisition unit 101 may acquire the prepared input information selected by the user from the first prepared input information, the second prepared input information, and the third prepared input information.

[0094] In the first modification example, we take the case where past content is stored in the user database DB1. Therefore, the input information acquisition unit 101 acquires the past content from the user database DB1. The past content may also be stored in a database other than the user database DB1, a computer other than the server 10, or an information storage medium. In this case, the input information acquisition unit 101 can acquire the past content from the other database, the other computer, or the information storage medium. The past content may also be associated with the user to whom the past content was provided, or the user who instructed the generation of the past content. When content C is generated to be provided to a user, or when a user instructs the generation of content C, the input information acquisition unit 101 may acquire the past content associated with these users.

[0095] Figure 9 shows an example of the input and output of the generative model in Modification Example 1. As shown in Figure 9, the prepared input information in Modification Example 1 consists of text instructing the generative model to generate content C that corresponds to past content. In the example in Figure 9, the prepared input information indicates that content C will be generated that has different characteristics from the past content. That is, the prepared input information indicates that past content should be excluded. For example, the prepared input information might be text such as, "You are an AI that generates content when sending electronic money. Please generate content that has different characteristics from the attached past content."

[0096] In the example shown in Figure 9, similar to the embodiment, the input information acquisition unit 101 acquires a prompt that includes user input information and prepared input information. In the modified example 1, the input information acquisition unit 101 may acquire the prepared input information directly as a prompt without acquiring user input information. That is, in modified example 1, user input information may not be acquired. The mode in which prepared input information is acquired without user input information being acquired in modified example 1 is also within the scope of this disclosure. The content provision system 1 in modified example 1 may not include a function to acquire user input information.

[0097] Furthermore, the generative model in Modification 1 is assumed to be a multimodal model capable of processing both the text indicated by the prompt and past content, which is in a data format different from text. A multimodal model is a model capable of processing multiple data formats. The generative model is assumed to have been trained on multiple formats of training data. The multimodal model itself may have a mechanism similar to that of known models. If the past content is in image format, the generative model is a model capable of processing both text and images.

[0098] For example, the content generation unit 102 inputs prompts and past content acquired by the input information acquisition unit 101 to the generation model. If the external system 2 manages the generation model, the content generation unit 102 inputs prompts and past content to the generation model by sending prompts and past content to the external system 2. If the generation model is stored in the data storage unit 100, the content generation unit 102 only needs to input prompts and past content to the generation model stored in the data storage unit 100.

[0099] For example, the generative model calculates an embedding representation corresponding to the prompt and past content based on parameters adjusted through pretraining, and outputs content C corresponding to the embedding representation. As shown in Figure 9, if the prepared input information indicates that content C should be generated which has different features from the past content, the generative model recognizes, based on the embedding representation of the prepared input information portion of the prompt, that it should generate content C which has different features from the past content input to itself, and generates content C which has different features from those indicated by the embedding representation of the past content input to itself. In the example in Figure 9, the flower shown in content C is larger in size and has a different shape than the flower shown in the past content, so it differs from the features of the past content.

[0100] Furthermore, if the input information indicates that the generation model should generate content C identical to past content, the generation model recognizes that it should generate content C identical to the past content input to it, and generates content C identical to the past content input to it. If the input information indicates that the generation model should generate content C that is not identical to past content but shares common features with past content, the generation model recognizes that it should generate content C that shares common features with the past content input to it, and generates content C that shares common features with the past content input to it.

[0101] Furthermore, while Modification 1 differs from the embodiment in that the content indicated by the prepared input information and past content are input to the generation model, the flow of processing executed after the generation model generates content C (the flow in which the content generation unit 102 acquires the content C generated by the generation model and the content provision unit 103 provides the content C) may be the same as in the embodiment.

[0102] In Modified Example 1, the content provision system 1 acquires prepared input information indicating that the generation model will generate content C corresponding to past content. This allows the content provision system 1 to provide the user with appropriate content C corresponding to past content. For example, if the prepared input information indicates that the generation model will generate content C with different characteristics from past content, the content provision system 1 can prevent the user from becoming bored by repeatedly being provided with content C similar to past content. If the prepared input information indicates that the generation model will generate content C with characteristics common to past content that the user liked, the content provision system 1 can provide content C with characteristics that the user liked.

[0103] [6-2. Variation 2] For example, in the embodiment, the content of Content C entered in input form F22 was described as an example of user input information. User input information can be any information entered by the user and is not limited to the example in the embodiment. Modification 2 describes other examples of user input information. In the example of input screen SC2 in the upper right and lower left of Figure 2, the user input information may be the recipient user specified by the sender user, the amount entered in input form F20, the message entered in input form F21, or a combination thereof. The user input information may show both the content of Content C described in the embodiment and the content described in Modification 2.

[0104] Figure 10 shows an example of the input and output of the generative model in Modification 2. The input information acquisition unit 101 in Modification 2 acquires a prompt based on at least one of the following: user identification information that can identify at least one of a plurality of users, a message to that at least one person, and content related to electronic money. For example, the user input information may show at least one of these, and the input information acquisition unit 101 may acquire a prompt based on the user input information showing at least one of these. At least one of the user identification information, message, and content related to electronic money does not have to be entered by the user and may be prepared in advance. In the example in Figure 10, the case where the user input information shows all of these is shown, but the user input information may show only one of these, or only two of these.

[0105] For example, when a sending user sends electronic money to a receiving user, if user input information is entered, the user input information will indicate the user identification information of the receiving user specified by the sending user. The user input information may also indicate the user identification information of multiple receiving users. When a receiving user requests electronic money from a sending user, if user input information is entered, the user input information will indicate the user identification information of the sending user specified by the receiving user. The user input information may also indicate the user identification information of multiple sending users.

[0106] For example, when a sending user sends electronic money to a receiving user, if they enter a message for the receiving user, the user input information will show the message entered by the sending user. The user input information may also show messages for each of multiple receiving users. When a receiving user requests electronic money from a sending user, if they enter a message for the sending user, the user input information will show the message entered by the receiving user. The user input information may also show messages for each of multiple sending users.

[0107] In the example in Figure 10, the amount of electronic money corresponds to the details related to the electronic money. The details related to the electronic money may also be other details such as the receipt deadline or payment deadline. If the payment service supports multiple types of electronic money, the details related to the electronic money may indicate the type of electronic money in question. When a sending user sends electronic money to a receiving user, the user input information indicates the details of the electronic money entered by the sending user. The user input information may also indicate the details of the electronic money sent to each of the multiple receiving users. When a receiving user requests electronic money from a sending user, the user input information indicates the details of the electronic money entered by the receiving user. The user input information may also indicate the details of the electronic money requested to each of the multiple sending users.

[0108] In the example shown in Figure 10, similar to the embodiment, the input information acquisition unit 101 acquires a prompt that includes user input information and prepared input information. For example, the prepared input information displays text such as, "You are an AI that generates content when sending electronic money. Please generate appropriate content based on the following user input information." This text may be the same as in Figure 6, but in the example shown in Figure 10, the content of the user input information set in the prepared input information is different.

[0109] In addition, the input information acquisition unit 101 in Modification 2 may acquire user input information directly as a prompt without including prepared input information. That is, in Modification 2, prepared input information does not need to be acquired. The mode in which user input information is acquired in Modification 2 without the acquisition of prepared input information is also within the scope of this disclosure. The content provision system 1 in Modification 2 does not need to include a function to acquire prepared input information.

[0110] For example, the content generation unit 102 inputs a prompt to the generation model in the same manner as in the embodiment. Based on the parameters adjusted by pretraining, the generation model calculates an embedded representation corresponding to the prompt and outputs content C corresponding to the embedded representation. As shown in Figure 10, if the prepared input information indicates that content C should be generated according to user input information, the generation model recognizes from the embedded representation of the prepared input information portion of the prompt that it should generate content C corresponding to the user input information entered into it, and generates content C corresponding to the embedded representation of the user input information entered into it.

[0111] For example, suppose the sending user enters a message such as, "This is for lunch the other day. Thank you." In this case, the generative model, based on the embedded expression of the word "lunch," generates an image of a typical lunch meal as content C, as shown in Figure 10. Alternatively, the generative model may recognize that the message is a thank you message based on the embedded expression of the word "thank you" and generate an image of flowers, which is appropriate as a thank you, as content C. Alternatively, the generative model may generate content C according to the amount of money included in the input information. The generative model may generate content C showing a more luxurious lunch or flowers the larger the amount.

[0112] In Modification 2, the prompt acquisition process differs from that of the embodiment, but the processing flow executed after the generation model generates content C (the content generation unit 102 acquires the content C generated by the generation model, and the content provision unit 103 provides the content C) may be the same as in the embodiment.

[0113] The content provision system 1 of the modified example 2 obtains a prompt based on at least one of the following: user identification information that can identify at least one of a group of users, a message to that at least one person, and the content of the electronic money. This allows the content provision system 1 to provide the user with appropriate content C according to the user input information. For example, the content provision system 1 can generate content C according to at least one of the sending user and the receiving user. The content provision system 1 can generate content C according to a message to at least one of the sending user and the receiving user. The content provision system 1 can generate content C according to the content of the electronic money.

[0114] [6-3. Modification 3] For example, in the embodiment, the case where user input information entered on input screen SC2 is input to the generation model was given as an example, but information that is not entered on input screen SC2 may also be acquired as input information. Hereafter, information that is not entered on input screen SC2 will be referred to as screen non-input information. Screen non-input information may not be information entered by the user, but may be pre-prepared information or information acquired on the spot. For example, screen non-input information may be basic user information, usage history information, the date and time at that time, the time of day at that time, or the season at that time.

[0115] Figure 11 shows an example of the input and output of the generation model in Modification Example 3. The input information acquisition unit 101 in Modification Example 3 acquires screen non-input information that is not entered on the input screen, relating to at least one of multiple users, and acquires a prompt based on the screen non-input information. The input information acquisition unit 101 may also acquire screen non-input information relating to the recipient user. For example, the input information acquisition unit 101 may acquire at least one of the recipient user's basic user information and usage history information as screen non-input information. The input information acquisition unit 101 may also acquire screen non-input information relating to the sender user. For example, the input information acquisition unit 101 may acquire at least one of the sender user's basic user information and usage history information as screen non-input information.

[0116] In variation 3, we take the case where screen non-input information is stored in the user database DB1 as an example. Therefore, the input information acquisition unit 101 acquires screen non-input information from the user database DB1. Screen non-input information may also be stored in a database other than the user database DB1, a computer other than the server 10, or an information storage medium. In this case, the input information acquisition unit 101 can acquire the screen non-input information from the other database, the other computer, or the information storage medium.

[0117] In the example in Figure 11, the prepared input information shows text such as, "You are an AI that generates content when sending electronic money. Please generate appropriate content based on the following screen-uninputted information." In Modification 3, the input information acquisition unit 101 may acquire a prompt based on the prepared input information and screen-uninputted information without acquiring user input information. That is, in Modification 3, user input information does not need to be acquired. The mode in which screen-uninputted information is acquired without acquiring user input information in Modification 3 is also within the scope of this disclosure. The content provision system 1 in Modification 3 does not need to include a function to acquire user input information.

[0118] For example, the content generation unit 102 inputs a prompt to the generation model in the same manner as in the embodiment. Based on the parameters adjusted by pre-training, the generation model calculates an embedded representation corresponding to the prompt and outputs content C corresponding to the embedded representation. As shown in Figure 11, if the prepared input information indicates that content C should be generated according to the non-input information on the screen, the generation model recognizes that it should generate content C corresponding to the non-input information input to itself based on the embedded representation of the non-input information portion of the prompt, and generates content C corresponding to the embedded representation of the non-input information input to itself. For example, if the non-input information on the screen indicates that the recipient user's preference is flowers, the generation model generates content C of flowers, as shown in Figure 11.

[0119] For example, suppose the recipient user's purchase history information shows that flowers were among the items the recipient user had previously purchased. In this case, the generative model recognizes that flowers are shown in the purchase history information based on the embedded representation of the purchase history information and generates content C that shows flowers. Alternatively, the generative model may generate content C according to the recipient user's age or age group as indicated in the input information. The generative model may also generate content C according to the recipient user's address as indicated in the input information. The generative model may also generate content C according to the current season as indicated in the input information.

[0120] In Modification 3, although it differs from the embodiment in that the prompt includes information about screen non-input, the processing flow executed after the generation model generates content C (the flow in which the content generation unit 102 acquires the content C generated by the generation model and the content provision unit 103 provides said content C) may be the same as in the embodiment.

[0121] The content delivery system 1 in Modification 3 acquires screen non-input information for at least one of multiple users and obtains a prompt based on the screen non-input information. This allows the content delivery system 1 to provide the user with appropriate content C according to the screen non-input information. For example, the content delivery system 1 can generate content C according to the attributes of at least one of the sending user and the receiving user. The content delivery system 1 can generate content C according to the usage history of the payment service by at least one of the sending user and the receiving user.

[0122] [6-4. Modification 4] For example, approval may be given for the provision of content C generated by the generative model. The content provision system 1 of Modification 4 includes an approval request unit 104. The approval request unit 104 requests approval for the provision of content C from at least one of the multiple users. Approval for the provision of content C is approval for the content provision unit 103 to provide content C. Approval can also be called permission. The approval request unit 104 requests approval from at least one user terminal 20 of the multiple users by sending display data for a screen to accept approval.

[0123] The user who is requested to approve may be the same user to whom content C is provided, or a user to whom content C is not provided. The approval request unit 104 may request approval from all of the users, or from some of the users. In the modified example 4, the approval request unit 104 requests approval from the sending user, but the approval request unit 104 may also request approval from the receiving user. The user database DB1 may store a setting indicating whether or not each user has given approval. Once a user has given approval, that setting may be referenced thereafter to provide content C.

[0124] Figure 12 shows an example of the input screen SC2 in Modification 4. In Modification 4, the example shows that approval is performed on the input screen SC2, but approval may be performed on a screen other than the input screen SC2. For example, approval may be performed on a screen before the input screen SC2 is displayed, on the confirmation screen SC3, on a screen displayed after the confirmation screen SC3, or on another screen. In the example in Figure 12, the input screen SC2 includes a checkbox C24 that accepts the approval operation. The sending user approves by checking the checkbox C24.

[0125] For example, when button B23 is selected, the user terminal 20 sends data to the server 10 indicating the selection result of checkbox C24. That is, the user terminal 20 sends data indicating whether or not approval has been granted. By receiving this data, the server 10 can determine whether or not approval has been granted. The content generation unit 102 may cause the generation model to generate content C if approval has been granted. That is, the content generation unit 102 may cause the generation model to generate content C on the condition that approval has been granted. The content generation unit 102 does not have to cause the generation model to generate content C if approval has not been granted.

[0126] In the modified example 4, the content provider unit 103 provides content C generated by the content generation unit 102 to at least one of multiple users if approval is granted. That is, the content provider unit 103 provides content C on the condition that approval is granted. If approval is not granted, the content provider unit 103 does not provide content C. Content C is generated by the content generation unit 102 regardless of whether approval is granted, and the provision of content C by the content provider unit 103 may be controlled based on whether approval is granted.

[0127] Approval may be performed by any operation. Approval is not limited to operations on checkbox C24. For example, approval may be performed by operations on other parts (parts as part of the user interface) other than checkbox C24. Other parts may be buttons such as radio buttons, parts equivalent to input forms other than buttons, text displayed on the screen, sliders, or icons. Server 10 can determine whether approval has been performed by obtaining data from user terminal 20 indicating the results of these arbitrary operations.

[0128] In the modified example 4, the content provision system 1 requests approval from at least one of the multiple users regarding the provision of content C. If approval is granted, the content provision system 1 provides content C, generated by the content generation unit 102, to at least one of the multiple users. This allows the content provision system 1 to provide content C only when approval is granted, thereby improving user convenience.

[0129] [6-5. Variation 5] For example, the input information acquisition unit 101 may acquire past payment information relating to electronic payments previously performed in the payment service, specifically past payment information for at least one of multiple users, as input information. Past payment information is the payment history from the usage history shown in the usage history information. Past payment information may show the payment history for the entire past period, or it may show the payment history for a specific period in the past.

[0130] In Modification 5, the example given is that past payment information indicates the most recent payment, but past payment information may also indicate the history of multiple payments. The input information acquisition unit 101 acquires information indicating the most recent payment from the usage history information stored in the user database DB1 as past payment information. Past payment information may also be stored in a database other than the user database DB1, a computer other than the server 10, or an information storage medium. The input information acquisition unit 101 may acquire past payment information from a database, a computer, or an information storage medium.

[0131] Figure 13 shows an example of the input and output of the generation model in Modification 5. In the example in Figure 13, the input information acquisition unit 101 of Modification 5 acquires a prompt that includes past payment information. The prepared input information of Modification 5 indicates an instruction to the generation model to generate content C according to the past payment information. For example, the prepared input information shows text such as, "You are an AI that generates content when sending electronic money. Please generate appropriate content based on the following past payment information."

[0132] Note that "[Past Payment Information]" in Figure 13 is a tag indicating the location where past payment information is set. Therefore, the input information acquisition unit 101 obtains the final prompt by setting the past payment information in the portion of the prepared input information indicated by the tag. Setting past payment information means inserting past payment information. Past payment information may be input into the generation model separately from the prepared input information, rather than being incorporated into the prepared input information. The input information acquisition unit 101 may also acquire the past payment information directly as a prompt without acquiring the prepared input information.

[0133] In Modification 5, the content generation unit 102 causes the generation model to generate content C based on past payment information. For example, the content generation unit 102 inputs a prompt to the generation model in the same manner as in the embodiment. The generation model calculates an embedded representation corresponding to the prompt based on parameters adjusted by pre-training and outputs content C corresponding to the embedded representation. As shown in Figure 13, if the prepared input information indicates that content C should be generated according to past payment information, the generation model recognizes from the embedded representation of the past payment information portion of the prompt that it should generate content C corresponding to the past payment information input to it, and generates content C corresponding to the embedded representation of the past payment information input to it. In the example in Figure 13, the past payment information indicates that flowers were purchased, and the generation model generates content C of flowers indicated by the past payment information.

[0134] In Modification 5, although it differs from the embodiment in that the prompt includes past payment information, the processing flow executed after the generation model generates content C (the flow in which the content generation unit 102 acquires the content C generated by the generation model and the content provision unit 103 provides the content C) may be the same as in the embodiment.

[0135] The content provision system 1 in Modification 5 acquires past payment information of at least one of multiple users as input information. Based on the past payment information, the content provision system 1 causes a generation model to generate content C. This allows the content provision system 1 to provide users with appropriate content C according to the past payment information. For example, the content provision system 1 can provide content C according to the most recent payment made by the sending user or the receiving user.

[0136] [6-6. Variation 6] For example, generating content C using a generative model can take some time. Using the flow shown in Figure 2 as an example, suppose the generative model generates content C between the time the sending user enters the necessary information into the input screen SC2 and the time the confirmation screen SC3 is displayed. In this case, if generating content C takes time, the display of the confirmation screen SC3 will also take time, which may cause the sending user to feel uneasy. Therefore, in Modification 6, we will consider an example where the content provision system 1 generates content C in advance to shorten the time required to display the confirmation screen SC3.

[0137] In the modified example 6, the input information acquisition unit 101 acquires input information before any of the multiple users perform a predetermined operation to send or request electronic money. The predetermined operation is an operation that is a condition for sending or requesting electronic money. The predetermined operation can be any operation that is predetermined. The input screen SC2 in the modified example 6 is given as an example where it does not include the input form F22. The sending user does not enter user input information on the input screen SC2. The sending user may enter user input information before the input screen SC2 is displayed.

[0138] For example, the predetermined operation may be an operation in which the sending user enters the necessary information for sending electronic money (e.g., the recipient user's user identification information, amount, message, content C, or a combination thereof), or it may be an operation performed afterward (e.g., selecting button B23). The predetermined operation may be an operation in which the recipient user enters the necessary information for requesting electronic money (e.g., the sending user's user identification information, amount, message, content C, or a combination thereof), or it may be an operation performed afterward.

[0139] For example, the input information acquisition unit 101 acquires input information when predetermined conditions are met. The predetermined conditions are conditions that trigger the input information acquisition unit 101 to acquire input information. For example, the predetermined conditions may be that the payment application is launched, a predetermined date and time arrives, the remittance function is selected in the payment application, an operation for displaying the input screen SC2 is performed (for example, selecting button B10), the user agrees to the AI ​​generating content (for example, the function for generating content with AI may be turnable on / off, and the user turns on that function), another operation is performed, or the remittance reaches a predetermined amount or predetermined number of times. The input information in Modification 6 may be any of the input information described in Embodiment and Modifications 1 to 5.

[0140] In Modification 6, an example is given where the arrival of a predetermined time each day constitutes a predetermined condition. Another example is given where the usage history information of the sending user constitutes the input information. For example, the input information acquisition unit 101 determines whether the predetermined condition has been met by identifying the current date and time using a real-time clock or a known means such as GPS, and determining whether the time corresponding to the predetermined condition has arrived. If the input information acquisition unit 101 determines that the predetermined condition has not been met, it does not acquire the input information. If the input information acquisition unit 101 determines that the predetermined condition has been met, it acquires the input information. For example, the input information acquisition unit 101 acquires the usage history information of multiple users from the user database DB1 as input information.

[0141] In the modified example 6, the content generation unit 102 causes the generation model to generate content C before any of the multiple users perform a predetermined operation. When predetermined conditions are met, the content generation unit 102 inputs the input information obtained for each user's content C into the generation model and causes the generation model to generate the user's content C. For each user, the content generation unit 102 stores the user's content C in the user database DB1, associated with the user's user ID and login account.

[0142] In Modification 6, the content provision unit 103 provides content C generated by the content generation unit 102 to at least one of the users after one of the users has performed a predetermined operation. When a user sends or requests electronic money, the content provision unit 103 retrieves the user's content C stored in the user database DB1. Although the content provision unit 103 differs from the embodiment and Modifications 1-5 in that it provides pre-generated content C, the processing for providing content C is the same as in the embodiment and Modifications 1-5.

[0143] In the modified version 6, the content provision system 1 acquires input information before any of the multiple users perform a predetermined operation to send or request electronic money. The content provision system 1 causes the generation model to generate content C before any of the multiple users perform the predetermined operation. After any of the multiple users perform the predetermined operation, the content provision system 1 provides the content C generated by the content generation unit 102 to at least one of the multiple users. This allows the content provision system 1 to reduce the time required to provide content C. For example, the content provision system 1 can reduce the time from when the sending user finishes inputting on the input screen SC2 until the confirmation screen SC3 is displayed.

[0144] [6-7. Variation 7] For example, the content generation unit 102 may cause the generation model to generate multiple content Cs. The number of content Cs to be generated by the generation model may be input to the generation model as a prompt, predetermined in the program of the generation model, or specified as a setting of the generation model. The number of content Cs to be generated by the generation model may be any number. For example, the number of content Cs to be generated by the generation model may be a specific number (for example, two, three, or four or more predetermined numbers), or it may be a numerical range (for example, a range such as two to five).

[0145] Figure 14 shows an example of the input and output of the generative model in Modification 7. Modification 7 takes the example of a case where the number of content C to be generated by the generative model is indicated in the prepared input information and input to the generative model as a prompt. For example, the prepared input information indicates the number of content C to be generated by the teacher model, such as "You are an AI that generates content when sending electronic money. Based on the input information you have entered, please generate three appropriate content items." The input information acquisition unit 101 in Modification 7 acquires a prompt that includes prepared input information indicating the number of content C to be generated by the generative model as input information.

[0146] In the modified example 7, the content generation unit 102 inputs a prompt to the generation model indicating the number of content C to be generated by the generation model. Based on parameters adjusted by pre-training, the generation model calculates an embedding representation corresponding to the prompt and outputs content C corresponding to the embedding representation. As shown in Figure 14, if the number of content C to be generated by the generation model is indicated in the prepared input information, the generation model recognizes the number of content C to be generated by the embedding representation of the part of the prompt that indicates the number of content C indicated in the prepared input information, and generates that number of content C based on the embedding representation of the other parts.

[0147] The number of content C generated by the generation model may be a predetermined fixed value, or it may be specified by the user. For example, the input screen SC2 may accept input for the number of content C generated by the generation model. In this case, the user input information will indicate the number of content C generated by the generation model. The number of content C may be entered in input form F22, or it may be entered in a separate dropdown menu.

[0148] In Modification 7, the content provider 103 provides at least one of several users with a content C selected by that at least one user or another user from among several content Cs. For example, each of the multiple content Cs generated as shown in Figure 14 is displayed on the confirmation screen SC3 or another screen as a candidate for the content C to be ultimately provided. The sending user selects one of the multiple content Cs. The content provider 103 provides the sending user with the confirmation screen SC3 containing the content C selected by the sending user. The content provider 103 provides the receiving user with the receiving user with the receiving user's content C containing the content C selected by the sending user. Alternatively, each of the multiple content Cs generated as shown in Figure 14 may be displayed on the confirmation screen SC3 or another screen as a candidate for the content C to be ultimately provided. Subsequently, electronic money is sent from the sending user to the receiving user. The receiving user may select one of the multiple content Cs before the receiving user displays the receiving screen SC4 to receive the electronic money. The content provision unit 103 may provide the recipient user with a receiving screen SC4 that includes the content C selected by the recipient user.

[0149] In the modified example 7, the content provision system 1 causes the generation model to generate multiple content Cs. The content provision system 1 provides at least one of the multiple users with a content C selected by that at least one user or another user from among the multiple content Cs. This allows the content provision system 1 to provide a more appropriate content C.

[0150] [6-8. Variation 8] For example, in the embodiment, a general-purpose generative AI capable of generating content C other than content C in the payment service is used as the generative model. In Modification 8, an example is given in which a model trained with training data specific to the payment service corresponds to the generative model. That is, the generative model in Modification 8 is a dedicated model specialized for generating content C in the payment service. In Modification 8, an example is given in which the data storage unit 100 stores the generative model before training. The generative model before training is a model with initial parameters. The generative model before training may be a model that has undergone some degree of pre-training. The model before training may be stored in an external system 2.

[0151] The content provision system 1 in Modification 8 includes a learning unit 105. The learning unit 105 trains a generative model with training data relating to the relationship between training input information, which is training input information, and training content corresponding to said training input information. Training data is data to be trained on the generative model. Training data is a pair of training input information and training content. Modification 8 takes the example of a case where multiple training data are stored in the training database DB2, but the multiple training data may be stored in a database other than the training database DB2, a computer other than the server 10, or an information storage medium. The learning unit 105 can obtain the training data from the training database DB2, another database, another computer, or an information storage medium. The training data is prepared by the administrator of the content provision system 1.

[0152] Training input information is information that is input to the generative model during training. Training input information has the same format as the input information that is input to the generative model during estimation. Modification 8 gives an example where the training input information includes training user basic information and training usage history information. Either training user basic information or training usage history information may constitute the training input information. Training input information may also be other information that has been described as an example of input information in Embodiment and Modifications 1 to 7, and which is other information prepared for training. Training content is content C that will be the correct answer during training. In other words, training content is content C that the generative model should ideally generate when training input information is input.

[0153] Figure 15 shows an example of the input and output of the generative model in Modification 8. An example of training data is also shown in Figure 15. The learning unit 105 performs training of the generative model based on the training data. The learning unit 105 adjusts the parameters of the generative model so that the generative model generates training content when training input information is input to the generative model. The training itself can utilize known training methods. For example, the learning unit 105 may perform training of the generative model based on backpropagation or gradient descent. The learning unit 105 may repeat training of the generative model based on the training data until the loss calculated by a predetermined loss function becomes small enough. When training is complete, the learning unit 105 overwrites and saves the generative model stored in the data storage unit 100.

[0154] In Modification 8, the content generation unit 102 causes the generation model, which has been learned by the learning unit 105, to generate content C. Although it differs from the embodiment and Modifications 1-7 in that learning is performed by the learning unit 105, the flow of content C generation is the same as in the embodiment and Modifications 1-7. In the example in Figure 15, the content generation unit 102 inputs the basic user information and usage history information of at least one of the source user and destination user as input information to the generation model. Based on the parameters adjusted by the learning unit 105, the generation model calculates the embedding representation of the input information and outputs content C corresponding to the embedding representation. The flow in which the content provision unit 103 provides the content C generated by the content generation unit 102 is also the same as in the embodiment and Modifications 1-7.

[0155] In the modified version 8, the content provision system 1 trains a generative model with training data relating to training input information and training content corresponding to the training input information. The content provision system 1 then has the generative model trained by the learning unit 105 generate content C. As a result, the content provision system 1 can create a generative model capable of generating content C specific to payment services, thereby improving the accuracy of content C.

[0156] [6-9. Other variations] For example, the above variations may be combined.

[0157] For example, the functions described as being implemented on server 10 may be implemented on user terminal 20 or other computers. The functions described as being implemented on server 10 may be shared among multiple computers.

[0158] [7. Addendum] For example, content delivery system 1 can also be configured as follows:

[0159] (1) An input information acquisition unit that acquires input information to be input into a generative model capable of generating content provided in a service where items of monetary value are sent between multiple users, A content generation unit that causes the generation model to generate the content based on the input information, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this. (2) The aforementioned generation model is a generation AI (Artificial Intelligence) model capable of generating the aforementioned content. The input information acquisition unit acquires a prompt for instructing the generation model to generate the content as the input information, The content generation unit causes the generation model to generate the content based on the prompt. (1) The content delivery system described above. (3) The input information acquisition unit acquires user input information related to input received on an input screen for any of the multiple users to send or request something of monetary value, and acquires the prompt based on the user input information. (2) The content delivery system described above. (4) The input information acquisition unit acquires user input information that indicates the content entered on the input screen by any of the multiple users. (3) The content delivery system described above. (5) The input information acquisition unit acquires prepared input information relating to pre-prepared inputs, and further acquires the prompt based on the prepared input information. A content delivery system as described in any of (2) to (4). (6) The input information acquisition unit acquires the prepared input information indicating that the generation model will generate content in accordance with user input information received on an input screen for any of the multiple users to send or request something of monetary value. (5) The content delivery system described above. (7) The input information acquisition unit acquires the prepared input information indicating that the generation model will generate content corresponding to past content previously generated by the generation model. (5) or (6) The content delivery system described above. (8) The input information acquisition unit acquires the prompt based on at least one of the following: user identification information that can identify at least one of the plurality of users, a message to that at least one person, and content relating to something of monetary value. A content delivery system as described in any of (2) to (7). (9) The input information acquisition unit acquires screen non-input information that is not entered on the input screen, relating to at least one of the multiple users, and acquires the prompt based on the screen non-input information. A content delivery system as described in any of (2) to (8). (10) The content provision system further includes an approval request unit that requests approval from at least one of the multiple users regarding the provision of the content, The content provision unit, upon approval, provides the content generated by the content generation unit to at least one of the multiple users. A content delivery system as described in any of (1) to (9). (11) The aforementioned service is a payment service that provides electronic payment to each of the aforementioned multiple users, The input information acquisition unit acquires past payment information relating to the electronic payment previously executed in the payment service, specifically the past payment information of at least one of the multiple users, as input information. The content generation unit causes the generation model to generate the content based on the past payment information. A content delivery system as described in any of (1) to (10). (12) The input information acquisition unit acquires the input information before any of the multiple users perform a predetermined operation to send or request the item of monetary value. The content generation unit causes the generation model to generate the content before any of the multiple users perform the predetermined operation. The content provision unit provides the content generated by the content generation unit to at least one of the users after any of the users has performed the predetermined operation. A content delivery system as described in any of (1) to (11). (13) The content generation unit causes the generation model to generate a plurality of the contents, The content provision unit provides to at least one of the multiple users the content selected by that at least one or other users from among the multiple content. A content delivery system as described in any of (1) to (12). (14) The content provision system further includes a learning unit that causes the generative model to learn training data relating to the relationship between training input information, which is the input information for training, and training content corresponding to the training input information. The content generation unit causes the generation model learned by the learning unit to generate the content. A content delivery system as described in any of (1) to (13). [Explanation of Symbols]

[0160] 1 Content provision system, 2 External system, 10 Server, 11,21 Control unit, 12,22 Storage unit, 13,23 Communication unit, 20 User terminal, 24 Operation unit, 25 Display unit, 100 Data storage unit, 101 Input information acquisition unit, 102 Content generation unit, 103 Content provision unit, 104 Approval request unit, 105 Learning unit, 200 Data storage unit, 201 Operation reception unit, 202 Display control unit, C Content, N Network, B10,B23,B40 Button, B30 Slide bar, C24 Check box, DB1 User database, DB2 Training database, F20,F21,F22 Input form, SC1 Top screen, SC2 Input screen, SC3 Confirmation screen, SC4 Receipt screen.

Claims

1. An input screen for any of several users to send or request something of monetary value, the input screen accepts input of an amount and input of a message to other users, and the input information acquisition unit acquires user input information relating to the input received on the input screen, and based on the user input information acquires a prompt to instruct the generation model to generate content, which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which something of monetary value is sent between the several users, as input information to be input to the generation model, A content generation unit that causes the generation model to generate the content based on the prompt, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this.

2. The input information acquisition unit acquires user input information that indicates the content entered on the input screen by any of the multiple users. The content provision system according to claim 1.

3. A device that acquires prepared input information relating to pre-prepared inputs for a generation model, which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which items of monetary value are sent between multiple users, the prepared input information indicating that the generation model will generate content in accordance with user input information received on an input screen that accepts input of an amount for which one of the multiple users will send or request the items of monetary value, and input of a message to the other users, and acquires an input information acquisition unit that acquires a prompt to instruct the generation model to generate the content based on the prepared input information, A content generation unit that causes the generation model to generate the content based on the prompt, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this.

4. A pre-prepared input information relating to a generation model, which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which items of monetary value are sent between multiple users, wherein the input information acquisition unit acquires the pre-prepared input information relating to pre-prepared inputs for a generation model, which is a generation AI (Artificial Intelligence) model capable of generating content provided in the service, wherein the generation model acquires the pre-prepared input information indicating that the generation model will generate content corresponding to the past content associated with the user to whom past content generated by the generation model was provided in the past, or the user who instructed the generation of the past content, and acquires a prompt to instruct the generation model to generate the content based on the pre-prepared input information. A content generation unit that causes the generation model to generate the content based on the prompt, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this.

5. An input information acquisition unit that acquires a prompt to instruct the generation model to generate content based on at least one of the following as input information to a generation model, which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which items of monetary value are sent between multiple users: user identification information capable of identifying at least one of the multiple users, a message to that at least one person, and content relating to the items of monetary value; A content generation unit that causes the generation model to generate the content based on the prompt, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this.

6. An input information acquisition unit that acquires screen non-input information relating to at least one of the multiple users, which is screen non-input information that is not entered on an input screen for any of the multiple users to send or request the item of monetary value, and acquires a prompt to instruct the generation model to generate the content, as input information to be input to a generation model which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which items of monetary value are sent among the multiple users, based on the screen non-input information; A content generation unit that causes the generation model to generate the content based on the prompt, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this.

7. An input screen for any of multiple users to send or request something of monetary value, the input screen which accepts input of an amount and input of a message to the other users, the input screen which includes an approval request unit that requests at least one of the multiple users to approve the provision of content provided in a service in which something of monetary value is sent between the multiple users, An input information acquisition unit acquires input information to be input into a generation model capable of generating the content based on the input to the input screen, A content generation unit that causes the generation model to generate the content based on the input information, If the aforementioned approval is granted, the content provision unit provides the content generated by the content generation unit to at least one of the multiple users, A content delivery system that includes this.

8. An input information acquisition unit that acquires past payment information relating to electronic payments previously executed by the payment service, as input information to be input to a generation model capable of generating content provided by a payment service in which something of monetary value is sent between multiple users and electronic payments are provided to each of the multiple users, A content generation unit that causes the generation model to generate the content by inputting a prompt containing the aforementioned past payment information into the generation model, A content provision unit provides the content generated by the content generation unit to at least one of the aforementioned multiple users, A content delivery system that includes this.

9. An input information acquisition unit that acquires input information to be input into a generation model capable of generating content provided in a service in which items of monetary value are sent between multiple users, before any of the multiple users perform a predetermined operation to send or request the items of monetary value, A content generation unit causes the generation model to generate the content based on the input information, in response to the launch of the service app, the arrival of a predetermined date and time, the selection of the money transfer function in the service app, the display of the input screen, the consent to the generation of content being made, or the sending of the valuable items reaching a predetermined amount or predetermined number of times, before any of the aforementioned users perform the predetermined operation. A content provision unit provides the content generated by the content generation unit to at least one of the users after any of the users has performed the predetermined operation. A content delivery system that includes this.

10. The content generation unit causes the generation model to generate a plurality of the contents, The content provision unit provides at least one of the multiple users with the content selected by that at least one or another user from among the multiple content. A content provision system according to any one of claims 1 to 9.

11. The content provision system further includes a learning unit that trains the generative model with training data relating to the relationship between training input information, which is input information for training, and training content corresponding to said training input information. The content generation unit causes the generation model learned by the learning unit to generate the content. A content provision system according to any one of claims 1 to 9.

12. An input screen for any of multiple users to send or request something of monetary value, the input screen accepts input of an amount and input of a message to other users, and the input information acquisition step acquires user input information relating to the input received on the input screen, and based on the user input information acquires a prompt to instruct the generation model to generate content, which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which something of monetary value is sent between the multiple users, as input information to be input to the generation model, A content generation step in which the generation model generates the content based on the prompt, A content provision step which provides the content generated by the content generation step to at least one of the multiple users, Content delivery methods including those mentioned above.

13. An input screen for any of multiple users to send or request something of monetary value, wherein the input screen accepts input of an amount and input of a message to other users, and the input information acquisition unit acquires user input information relating to the input received on the input screen, and based on the user input information acquires a prompt to instruct the generation model to generate content, which is a generation AI (Artificial Intelligence) model capable of generating content provided in a service in which something of monetary value is sent between the multiple users, as input information to be input to the generation model, A content generation unit causes the generation model to generate the content based on the prompt. A content provision unit provides the content generated by the content generation unit to at least one of the multiple users. A program that makes a computer function.

Images