Artificial intelligence system and program of artificial intelligence system

Abstract

To provide an artificial intelligence system and a program that enable a user to obtain information on desired generated text data based on the prompt of predetermined command and instruction information for an uploaded document file.SOLUTION: An artificial intelligence system 100 includes a user terminal 110 and servers (120 and 130). When an operation of designating a document file DF at a browser 111 of a user terminal is given, the user terminal uploads the document file to the server. When a predetermined operation at the browser is given, the user terminal transmits, to the server, instruction and command information PT in accordance with the predetermined operation. The server generates, based on the instruction and command information, text data for the read document file using artificial intelligence means 131 that has already learnt, and transmits, to the user terminal, the obtained generated text data GT. The user terminal displays, in the browser, the generated text data GT.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to an artificial intelligence system that includes a user terminal and a server, in which the user terminal outputs a display or voice based on the server, and a program for the system. [Background technology]

[0002] 2. Description of the Related Art Conventionally, server systems that provide functions such as saving and displaying files uploaded by users are known (for example, see Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2023-66473 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, in the conventional server system described above, files uploaded to cloud storage could only be viewed, downloaded, and edited by users other than the user who uploaded the file, depending on the permissions granted to them. In order to know the contents of the file, however, the user had to open the file and read it himself / herself. If the file contained a large amount of information, such as many pages, this took time, making it difficult to grasp the contents to a certain extent in a short period of time.

[0005] Therefore, the present invention solves the problems of the prior art as described above. That is, the object of the present invention is to provide an artificial intelligence system and a program thereof that allows a user to obtain desired generated text data information based on prompts of specified command information for the uploaded document file, or the webmail page of the URL information, or the document file identified by the URL information, simply by uploading a document file or URL information and performing a specified operation. [Means for solving the problem]

[0006] The invention of claim 1 is an artificial intelligence system comprising a user terminal and a server, in which the user terminal displays or outputs audio based on the server, and when an operation to specify a document file or an operation to specify URL information which is location information for specified information is performed in the browser of the user terminal, the user terminal uploads the document file or URL information to the server, and when a specified operation is performed in the browser, the user terminal transmits instruction command information corresponding to the specified operation to the server, and the server uses a pre-trained artificial intelligence means based on the received instruction command information to generate text data for the content information identified by the loaded document file or URL information, and transmits the generated text data obtained by the execution to the user terminal, and the user terminal displays the received generated text data in the browser or outputs it as audio. This solves the above-mentioned problem.

[0007] The invention of claim 2 further solves the above-mentioned problem by being configured in such a way that, in addition to the configuration of the artificial intelligence system described in claim 1, when the pre-trained artificial intelligence means generates text data based on instruction command information, the pre-trained artificial intelligence means generates basis information regarding which page or part of the content information identified by the loaded document file or URL information the text data is based on, the server transmits the generated text data and basis information to a user terminal, and the user terminal displays the generated text data and basis information in a browser.

[0008] The invention of claim 3 further solves the above-mentioned problem by being configured in such a way that, in addition to the configuration of the artificial intelligence system described in claim 2, when the pre-trained artificial intelligence means generates text data based on instruction command information, the pre-trained artificial intelligence means generates multiple related question information in which information regarding the content information identified by the loaded document file or URL information serves as answer information, and the server transmits the generated text data and the generated question information to the user terminal in that order.

[0009] The invention of claim 4 further solves the above-mentioned problem by being configured in such a way that, in addition to the configuration of the artificial intelligence system described in claim 3, the instruction and command information is a question about content information identified by an uploaded document file or URL information, and the pre-trained artificial intelligence means calculates a correlation score between the content of the question and the content of the information about the content information identified by the loaded document file or URL information, and if the calculated correlation score is less than a predetermined threshold, the server sends generated text data to the user terminal indicating that the answer to the question is unknown.

[0010] The invention of claim 5 further solves the above-mentioned problem by being configured such that, in addition to the configuration of the artificial intelligence system described in claim 4, when the instruction command information includes language designation information for the answer, the pre-trained artificial intelligence means generates text data in a designated language based on the instruction command information, and the server transmits the generated text data to a user terminal, and when the specified operation is an instruction command input operation in a foreign language other than Japanese, the pre-trained artificial intelligence means generates text data in the language of the input instruction command based on the instruction command information, and the server transmits the generated text data to a user terminal.

[0011] The invention according to claim 6 further comprises, in addition to the configuration of the artificial intelligence system according to any one of claims 1 to 5, the user terminal communicates with a second server having pre-trained artificial intelligence means via a first server having a bot, and on an administration screen of an administrator terminal capable of communicating with the first server, masking of word attributes contained in data transmitted from the user terminal to the bot of the first server can be set freely, and when data including characters of a word attribute of a masking preset in a browser of the user terminal is input and uploaded to the first server, the bot of the first server that receives the data determines that the received data includes characters of a word attribute of the masking, and The above-mentioned problem is further solved by the configuration in which characters with a masking word attribute in the data are converted into other arbitrary characters with the same word attribute, and the converted data is then sent to the pre-trained artificial intelligence means of the second server, the second server uses the pre-trained artificial intelligence means of the second server to generate text data based on the converted data, and the generated text data obtained by this execution is sent to the bot of the first server, and the bot of the first server changes the part of the received generated text data that was converted into other arbitrary characters when sending it to the pre-trained artificial intelligence means of the second server back to the characters before conversion, and then sends the changed generated text data to the user terminal.

[0012] The invention of claim 7 further solves the above-mentioned problem by, in addition to the configuration of the artificial intelligence system described in claim 6, configuring the second server to use a pre-trained artificial intelligence means to generate text data for the content information identified by the loaded document file or URL information based on the received instruction command information to summarize, and to transmit the generated text data obtained by the execution to a user terminal via the first server.

[0013] The invention of claim 8 further solves the above-mentioned problem by being configured in such a way that, in addition to the configuration of the artificial intelligence system described in claim 6, the instruction command information is to check whether there are any problems and, if there are any problems, to explain the details of the problems, and the second server uses a pre-trained artificial intelligence means to generate text data for the content information identified by the loaded document file or URL information based on the received instruction command information to check whether there are any problems and, if there are any problems, to explain the details, and the generated text data obtained by the execution is transmitted to the user terminal via the first server.

[0014] The invention of claim 9 further solves the above-mentioned problem by being configured in addition to the artificial intelligence system configuration described in claim 6, such that when the number of document files or URL information that can be specified in the browser of the user terminal is multiple, and multiple document files or URL information are specified, the user terminal uploads the multiple document files or URL information to the first server.

[0015] The invention of claim 10 is a program for an artificial intelligence system in which a user terminal displays or outputs audio based on a server, the program comprising: a file designation determination step for determining whether an operation for designating a document file or an operation for designating URL information which is location information for specified information has been performed in the browser of the user terminal; an upload step for the user terminal to upload a document file or URL information to the server when the operation for designating the document file or the operation for designating URL information which is location information for specified information has been performed; a predetermined operation determination step for determining whether a predetermined operation has been performed in the browser; an instruction command transmission step for the user terminal to transmit instruction command information corresponding to the predetermined operation to the server when the predetermined operation has been performed; a text data generation execution step for the server to generate text data for content information specified by the loaded document file or URL information using a pre-trained artificial intelligence means based on the received instruction command information; a generated data transmission step for the server to transmit the generated text data obtained by the execution to the user terminal; and a display / audio output step for the user terminal to display or output the received generated text data in the browser or audio. Effect of the Invention

[0016] The artificial intelligence system of the present invention is equipped with a user terminal and a server, which not only enables the user terminal to display or output audio based on the server, but also provides the following unique effects.

[0017] According to the artificial intelligence system of the invention of claim 1, the pre-trained artificial intelligence means reads the content information identified by the uploaded document file or URL information, and then generates text data based on instruction command information corresponding to a specified operation, and the generated text data is displayed in a browser or output as voice. Therefore, a user can obtain desired generated text data information based on prompts of specified instruction command information, such as creating a summary, pointing out problems, pointing out areas for improvement, and questioning information, without having to read the uploaded document file himself / herself, or the webmail page of the uploaded URL information or the document file identified by the URL information, simply by uploading a document file or URL information and performing a specified operation.

[0018] According to the artificial intelligence system of the invention of claim 2, in addition to the effect achieved by the invention of claim 1, not only the generated text data but also the basis information is displayed in the browser, so that the user can easily check the accuracy and reliability (veracity) of the generated text data by referring to the content information identified by the original document file or URL information based on the basis information. In other words, the user can easily confirm which part and what the text data was generated based on, without simply looking at the generated text data and accepting the contents of the generated text data.

[0019] According to the artificial intelligence system of the invention of claim 3, in addition to the effects achieved by the invention of claim 2, multiple anticipated question information about the content information identified by the loaded document file or URL information is generated and displayed in the browser in the order following the generated text data, so that the user can first select one of the multiple question information and interact with the pre-trained artificial intelligence means to easily deepen his / her understanding of the content information identified by the document file or URL information loaded by the pre-trained artificial intelligence means without having to read all of it himself / herself.

[0020] According to the artificial intelligence system of the invention of claim 4, in addition to the effect of the invention of claim 3, when a question is asked about content that has a low correlation with the content information specified by the uploaded document file or URL information, an honest answer that the user does not know is displayed in the browser without giving an answer with a low reliability, so that the user can understand that the question he or she asked was irrelevant in relation to the content information specified by the uploaded document file or URL information and try a different question. This means that low-confidence content will not be displayed as answers in the browser, allowing users to interact with pre-trained artificial intelligence tools with confidence.

[0021] According to the artificial intelligence system of the invention of claim 5, in addition to the effect achieved by the invention of claim 4, generated text data in a specified language is displayed as an answer in a browser, making it possible to accommodate multilingual users. Furthermore, when a question or other instruction is entered in English or another language, text data is generated in that language and displayed in the browser as a response, making it possible to accommodate multilingual users.

[0022] According to the artificial intelligence system of the invention of claim 6, in addition to the effects of the invention of any one of claims 1 to 5, for example, by pre-setting masking of word attributes that can identify individuals, such as name, address, telephone number, and email address, for data from a user terminal to a bot on the first server, if the data entered by the user contains characters with these word attributes, the bot on the first server converts the characters with these word attributes into any other characters with the same word attributes and then transmits the converted data to the pre-trained artificial intelligence means of the second server, and text data generation is performed based on the converted data, thereby preventing the chat GPT, which is an example of the pre-trained artificial intelligence means of the second server, from memorizing information about characters with word attributes that can identify individuals. In other words, since the content sent to Chat GPT, which is an example of the second server's pre-trained artificial intelligence means, is converted into characters that cannot identify individuals, it is possible to avoid the leakage of personally identifiable information to other terminals via Chat GPT, which is an example of the second server's pre-trained artificial intelligence means. Furthermore, when the bot of the first server receives generated text data from the pre-trained artificial intelligence means of the second server, the part of the received generated text data that was converted into another arbitrary character when it was sent to the pre-trained artificial intelligence means of the second server is modified back to the character before conversion, and the modified generated text data is sent to the user terminal, so that modified generated text data that basically conforms to the intent of the instruction command information entered by the user can be obtained. In other words, the characters converted when the bot on the first server sends to the pre-trained artificial intelligence means of the second server are converted back to the original characters when the bot on the first server receives them from the pre-trained artificial intelligence means of the second server, thereby avoiding unintended information leakage to the pre-trained artificial intelligence means of the second server or to other terminals via the pre-trained artificial intelligence means of the second server, and ensuring consistency between the instructions and commands from the user terminal and the responses to the user terminal.

[0023] According to the artificial intelligence system of the invention of claim 7, in addition to the effect of the invention of claim 6, the user only needs to perform a predetermined operation corresponding to the instruction command to summarize by specifying a document file or URL information, and the instruction command information to summarize is transmitted to the second server via the first server and processed by the pre-trained artificial intelligence means, and the desired result information is displayed on the browser or output as voice, so that the user can obtain the result information of the desired summary generation. In particular, when the specified document file or URL information is a long text, a summary of a short text is displayed as the result information or output as voice, so that the user does not need to read and understand the entire text of the original document file or URL information, and can understand the general content by reading or listening to the short text.

[0024] According to the artificial intelligence system of the invention of claim 8, in addition to the effects of the invention of claim 6, the user simply needs to specify a document file or URL information, perform a specified operation corresponding to an instruction command to check whether there are any problems and, if there are any problems, explain the details thereof. The instruction command information to check whether there are any problems and, if there are any problems, explain the details thereof is sent to the second server via the first server, processed by the pre-trained artificial intelligence means, and the desired result information is displayed in the browser or output as audio, so that the user can obtain the result information of the desired explanatory text generation regarding the problems in the content of the specified document file or URL information without reading the content page of the specified document file or URL information.

[0025] According to the artificial intelligence system of the invention of claim 9, in addition to the effect achieved by the invention of claim 6, the contents of multiple document files or URL information are loaded into a pre-trained artificial intelligence means, so that result information can be obtained about the common content of the contents of the multiple document files or URL information or the content of at least one of them. In other words, it is possible to obtain efficient result information that takes into consideration commonalities and differences in the contents of a plurality of document files or URL information.

[0026] According to the program of the artificial intelligence system of the invention of claim 10, similar to the effect achieved by the invention of claim 1, the pre-trained artificial intelligence means reads the content information identified by the uploaded document file or URL information, and then generates text data based on instruction command information corresponding to a specified operation, and the generated text data is displayed in a browser or output as voice. Therefore, the user can obtain desired generated text data information based on prompts of specified instruction command information, such as creating a summary, pointing out problems, pointing out areas for improvement, and questioning information, without having to read the uploaded document file himself, or the webmail page of the uploaded URL information or the document file identified by the URL information, by simply uploading a document file or URL information and performing a specified operation. [Brief description of the drawings]

[0027] [Figure 1] FIG. 1 is a diagram showing the concept of an artificial intelligence system according to an embodiment of the present invention. [Diagram 2] FIG. 2 is a chart showing an example of the operation of the artificial intelligence system according to the embodiment of the present invention. [Diagram 3] FIG. 2 is a diagram showing an example of file designation on the upload screen of a browser on a user terminal of the artificial intelligence system according to an embodiment of the present invention. [Figure 4] FIG. 2 is a diagram showing an example of a prompt input operation on the upload screen of a browser of a user terminal of the artificial intelligence system according to an embodiment of the present invention. [Diagram 5] FIG. 13 is a diagram showing an example of generated text data as an answer on the upload screen of a browser on a user terminal of the artificial intelligence system according to an embodiment of the present invention. [Figure 6] A figure showing an example of generated text data and multiple question information as answers on the upload screen of a browser of a user terminal of an artificial intelligence system according to an embodiment of the present invention. [Figure 7]A figure showing an example of questions as prompts and generated text data as answers on the upload screen of a browser on a user terminal of an artificial intelligence system according to an embodiment of the present invention. [Figure 8] FIG. 13 shows an example of an answer instruction in English as a prompt on the upload screen of a browser on a user terminal of an artificial intelligence system according to an embodiment of the present invention, and an example of generated text data in English as an answer. [Figure 9] A figure showing an example of English questions as prompts and generated text data in English as answers on the upload screen of a browser of a user terminal of an artificial intelligence system according to an embodiment of the present invention. [Figure 10] FIG. 1 is a diagram showing the concept of a masking function of an artificial intelligence system according to an embodiment of the present invention. [Figure 11] 1A and 1B are diagrams showing an example of masking settings on the management screen of an artificial intelligence system that is an embodiment of the present invention, and a diagram showing what happens when a document file containing characters to be masked is uploaded / input on the upload screen. [Figure 12] 1A to 1C are diagrams illustrating an example of the masking operation of the artificial intelligence system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] The artificial intelligence system of the present invention comprises a user terminal and a server, and when an operation to specify a document file or an operation to specify URL information, which is location information for specified information, is performed in the browser of the user terminal, the user terminal uploads the document file or URL information to the server, and when a specified operation is performed in the browser, the user terminal transmits instruction command information corresponding to the specified operation to the server, and the server uses a pre-trained artificial intelligence means based on the received instruction command information to generate text data for the content information identified by the loaded document file or URL information, transmits the generated text data obtained by the execution to the user terminal, and the user terminal displays the received generated text data on the browser or outputs it as audio. Therefore, the specific embodiment of the system may be any one as long as the user can obtain desired generated text data information based on a prompt of specified instruction command information, such as, for example, creating a summary, pointing out problems, pointing out areas for improvement, or questioning information, without having to read the uploaded document file himself or herself, or without having to read the webmail page of the uploaded URL information or the document file identified by the URL information himself or herself. The program for the artificial intelligence system of the present invention includes a file designation determination step for determining whether or not an operation for designating a document file or an operation for designating URL information that is location information for the predetermined information has been performed in the browser of the user terminal; an upload step for the user terminal to upload the document file or the URL information to the server when the operation for designating the document file or the operation for designating URL information that is location information for the predetermined information has been performed; a predetermined operation determination step for determining whether or not a predetermined operation has been performed in the browser; an instruction command transmission step for the user terminal to transmit instruction command information corresponding to the predetermined operation to the server when the predetermined operation has been performed; and a command transmission step for the server to transmit instruction command information for the content information specified by the loaded document file or URL information using the pre-trained artificial intelligence means based on the received instruction command information. The specific implementation may be any as long as it has a text data generation execution step of generating text data, a generated data transmission step in which the server transmits the generated text data obtained by the execution to a user terminal, and a display / audio output step in which the user terminal displays the received generated text data in a browser or outputs it as audio, thereby enabling a user to obtain desired generated text data information based on prompts of specified command information, such as creating a summary, pointing out problems, pointing out areas for improvement, or questioning information, by simply uploading a document file or URL information and performing a specified operation, without having to read the uploaded document file himself / herself, or without having to read the webmail page of the uploaded URL information or the document file identified by the URL information himself / herself.

[0029] For example, the user terminal may be a desktop personal computer terminal, a notebook personal computer terminal, a smartphone terminal, a tablet terminal, or the like that sends and receives information, and may be any terminal that can be freely connected to a server via a communication network including a wide area network, such as the Internet, a local network, or a telephone line. Also, the server may be a single server or multiple servers on a cloud. The pre-trained artificial intelligence means may be composed of an interactive model, also known as a large-scale language model, such as ChatGPT (Generative Pre-trained Transformer) (hereinafter referred to as ChatGPT), and may be configured to be installed on one server or multiple servers on the cloud. In addition, a document file refers to an electronic data file whose contents include character information, and the data file format may be a simple text data format, a document data format, a cell-type spreadsheet file format, a pixel-type image file format such as a photo or video, etc., which allows characters to be read using an image recognition means or which allows characters to be read using a voice recognition means, or any format that allows characters to be directly recognized or indirectly recognized by converting voice to characters. The characters also include the source code of a program. Furthermore, the URL information (Uniform Resource Locator information), which is the location information for the specified information, may be any information that can identify a website page in HTML format or a document file as the specified information obtained by accessing a web server. EXAMPLES

[0030] An artificial intelligence system 100 according to an embodiment of the present invention will be described below with reference to FIGS. Here, FIG. 1 is a diagram showing the concept of an artificial intelligence system 100 according to an embodiment of the present invention, FIG. 2 is a chart diagram showing an example of the operation of the artificial intelligence system 100 according to an embodiment of the present invention, FIG. 3 is a diagram showing an example of file designation on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention, FIG. 4 is a diagram showing an example of a prompt input operation on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention, FIG. 5 is a diagram showing an example of generated text data GT as an answer on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention, FIG. 6 is a diagram showing an example of generated text data GT and a plurality of question information as answers on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention, and FIG. 7 is a diagram showing an example of a prompt on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention. FIG. 8 is a diagram showing an example of generated text data GT in English as a prompt for an answer instruction and in English as an answer on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention. FIG. 9 is a diagram showing an example of a question in English as a prompt and generated text data GT in English as an answer on the upload screen 112 of the browser 111 of the user terminal 110 of the artificial intelligence system 100 according to an embodiment of the present invention. FIG. 10 is a diagram showing the concept of a masking function of the artificial intelligence system 100 according to an embodiment of the present invention. FIG. 11(A) is a diagram showing an example of masking settings on the management screen 113 of the artificial intelligence system 100 according to an embodiment of the present invention. FIG. 11(B) is a diagram showing a state when a document file DF including characters to be masked is uploaded on the upload screen 112. FIG. 12 is a diagram explaining an example of the masking operation of the artificial intelligence system 100 according to an embodiment of the present invention.

[0031] An artificial intelligence system 100 according to a first embodiment of the present invention includes a user terminal 110, and a first server 120 and a second server 130, which are examples of servers, as shown in FIG. Among these, the first server 120 has information on the web page of the upload screen 112 of the artificial intelligence system 100 displayed by the browser 111 of the user terminal 110 . On the other hand, the second server 130 has a chat GPT, which is an example of a pre-trained artificial intelligence means 131. In this embodiment, as an example, the first server 120 and the second server 130 are configured as separate entities, but they may be configured as a single shared server. When a user performs a login operation to the artificial intelligence system 100 on the browser 111 of the user terminal 110 , the browser 111 of the user terminal 110 displays an upload screen 112 .

[0032] It is assumed that an operation to designate a document file DF or an operation to designate URL information DF1, which is location information for predetermined information, has been performed on upload screen 112 of browser 111 of user terminal 110. Then, the user terminal 110 uploads the document file DF or the URL information DF1 to the second server 130 via the first server 120, which is an example of a server. Next, when a predetermined operation is performed on upload screen 112 of browser 111, user terminal 110 transmits instruction command information PT, which is a so-called prompt in response to the predetermined operation, to second server 130 via first server 120, which is an example of a server. For example, it transmits predetermined command information such as making a summary, pointing out problems, pointing out areas for improvement, and questioning information.

[0033] Then, the second server 130 uses the chat GPT, which is an example of the pre-trained artificial intelligence means 131, based on the received instruction command information PT to generate text data for the content information identified by the loaded document file DF or URL information DF1. Then, the second server 130 transmits the generated text data GT obtained through the execution to the user terminal 110 via the first server 120. The user terminal 110 is configured to then display the received generated text data GT on the upload screen 112 of the browser 111 or output it as voice.

[0034] As a result, the chat GPT, which is an example of a pre-trained artificial intelligence means 131, reads the uploaded document file DF or the content information identified by the URL information DF1, and then generates text data based on the instruction command information PT corresponding to a specified operation, and the generated text data GT is displayed on the browser 111 or output as audio. As a result, the user can obtain desired information in generated text data GT based on prompts of specified command information, such as making a summary, pointing out problems, pointing out areas for improvement, and questioning information, by simply uploading a document file DF or URL information DF1 and performing a specified operation, without having to read the uploaded document file DF himself / herself, or without having to read the webmail page of the uploaded URL information DF1 or the document file DF identified by the URL information DF1 himself / herself.

[0035] Next, an example of the operation of the artificial intelligence system 100 will be described in more detail. As shown in FIG. 2, in step S1, which is a file designation determination step, the user terminal 110 determines whether or not an operation has been performed to designate a document file DF or an operation to designate URL information DF1, which is location information for specified information, on the upload screen 112 of the browser 111 of the user terminal 110. If it is determined that a designation operation has been performed, the process proceeds to step S2, whereas if it is determined that a designation operation has not been performed yet, step S1 is repeated.

[0036] More specifically, as shown in FIG. 3, upload screen 112 is provided with a document file specification field and a document file selection field. The document file specification field is provided so that when the icon of a document file DF is moved into the area of ​​the document file specification field by a drag-and-drop operation by the user, that document file DF is specified. The document file selection field is also provided so that a document file DF can be specified by selecting the path of the document file DF saved in the user terminal 110 or a file server by operating a reference button.

[0037] Subsequently, when the "upload" button is operated, it is determined that an operation to designate a document file DF or an operation to designate URL information DF1 which is location information for the predetermined information has been performed, and the process proceeds to step S2. It may also be determined that an operation to specify a document file DF has occurred when the icon of the document file DF is moved into the area of ​​the document file specification field by a drag-and-drop operation, or when the path of the document file DF is selected. In other words, it may be determined that an operation to specify the document file DF has been performed without operating the "upload" button. Also, as an example, the document file DF is designated by drag-and-drop in the document file designation field, but URL information DF1, which is location information for the predetermined information, may be input.

[0038] In step S2, which is an upload step, the user terminal 110 uploads the specified document file DF or URL information DF1 to the second server 130 via the first server 120, which is an example of a server. For example, when the user terminal 110 uploads a document file DF to the first server 120, the first server 120 converts the document file DF, such as a PDF file, into text, that is, into text data. Then, the first server 120 may transmit the text data to the second server 130. If the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, has a limit on the number of characters that can be input at one time, the first server 120 sends text data within the limited number of input characters. The same is true for the URL information DF1. When user terminal 110 uploads URL information DF1 to first server 120, first server 120 reads the contents of the specific information based on URL information DF1 to obtain text data. Then, the first server 120 transmits the text data to the second server 130. In step S3, as a predetermined operation determination step, the user terminal 110 determines whether or not a predetermined operation has been performed on the upload screen 112 of the browser 111. If it is determined that the predetermined operation has been performed, the process proceeds to step S4, and if it is determined that the predetermined operation has not been performed yet, step S3 is repeated. More specifically, as shown in FIG. 4, when data DT, which is an example of a specified operation, is input into a text input field provided on upload screen 112, that is, when instruction command information PT, for example, instruction command information PT to summarize, such as "↑ Summarize!", is input as a prompt, it is determined that a specified operation has been performed.

[0039] In addition, when instruction command information PT is operated not only directly in the text input field but also by operating a prompt button provided in the left column, instruction command information PT corresponding to the prompt button is input into the text input field and it is determined that a specified operation has been performed. Here, as an example of a specified operation, by operating the "Summarize!" prompt button, instruction command information PT to summarize the contents of document file DF is entered into a text input field, by operating the "Tell me the problems!" prompt button, instruction command information PT to point out problems with the contents of document file DF is entered into a text input field, and by operating the "Suggest improvements" prompt button, instruction command information PT to point out improvements with the contents of document file DF is entered into a text input field.

[0040] In step S4, which is an instruction / command sending step, as shown in FIG. 5, the user terminal 110 sends a prompt, "↑Summarize!", which is instruction / command information PT in response to a specified operation, to the second server 130 via the first server 120, which is an example of a server. In step S5, as a text data generation execution step, the second server 130, which is an example of a server, executes text data generation for the content information identified by the loaded document file DF or URL information DF1 using the chat GPT, which is an example of a pre-trained artificial intelligence means 131, based on the received instruction command information PT.

[0041] As a technical idea, the data DT entered in the text input field in the specified operation determination step S3 is transmitted to the second server 130 via the first server 120 in the instruction command transmission step S4, regardless of whether or not it includes instruction command information PT. Then, in the text data generation execution step S5, if the data DT received by the second server 130 includes instruction command information PT, the chat GPT, which is an example of the pre-trained artificial intelligence means 131, executes text data generation for the contents of the document file DF based on the instruction command information PT. To facilitate understanding of the present invention, an example will be described in which the data DT entered in the text input field in the predetermined operation determination step S3 contains instruction command information PT.

[0042] In step S6, as a generated data transmitting step, second server 130, which is an example of a server, transmits generated text data GT obtained by execution to user terminal 110 via first server 120. In step S7, which is a display / audio output step, the user terminal 110 displays the received generated text data GT on the upload screen 112 of the browser 111 or outputs the same as audio, as shown in FIG.

[0043] As a result, as described above, the chat GPT, which is an example of the pre-trained artificial intelligence means 131, reads the uploaded document file DF or the content information identified by the URL information DF1, and then generates text data based on the instruction command information PT corresponding to a specified operation, and the generated text data GT is displayed on the browser 111 or output as audio. As a result, the user can obtain desired information in generated text data GT based on prompts of specified command information, such as making a summary, pointing out problems, pointing out areas for improvement, and questioning information, by simply uploading a document file DF or URL information DF1 and performing a specified operation, without having to read the uploaded document file DF himself / herself, or without having to read the webmail page of the uploaded URL information DF1 or the document file DF identified by the URL information DF1 himself / herself.

[0044] For example, suppose that the instruction and command information PT is to summarize. In this case, the second server 130 uses the chat GPT, which is an example of the pre-trained artificial intelligence means 131, based on the received instruction command information PT to summarize, to generate text data for the content information identified by the loaded document file DF or URL information DF1. Then, the second server 130 transmits the generated text data GT obtained through the execution to the user terminal 110 via the first server 120.

[0045] As a result, the user simply designates the document file DF or URL information DF1 and performs a specified operation corresponding to the instruction command to summarize, and the instruction command information PT to summarize is sent to the second server 130 via the first server 120, processed by the chat GPT, which is an example of the pre-trained artificial intelligence means 131, and the desired result information is displayed on the upload screen 112 of the browser 111 or output as audio. As a result, the user can obtain the desired summary generation result information. In particular, when the specified document file DF or URL information DF1 is a long text, a summary of the short text is displayed as the result information or output as voice. As a result, the user does not need to read and understand the entire text of the original document file DF or URL information DF1, but can understand the general content by reading or listening to short sentences.

[0046] Also, for example, it is assumed that the instruction and command information PT is to check whether there are any problems and, if there are any problems, to explain the details thereof. In this case, the second server 130 checks whether there are any problems received, and if there are any problems, uses the chat GPT, which is an example of the pre-trained artificial intelligence means 131, based on the instruction command information PT to explain the contents of the problems, to generate text data for the content information identified by the loaded document file DF or URL information DF1. Then, the second server 130 transmits the generated text data GT obtained through the execution to the user terminal 110 via the first server 120.

[0047] As a result, the user simply designates a document file DF or URL information DF1 and performs a specified operation corresponding to an instruction command to check whether there are any problems and, if there are any problems, to explain the details thereof, and the instruction command information PT to check whether there are any problems and, if there are any problems, to explain the details thereof is transmitted to the second server 130 via the first server 120, where it is processed by the chat GPT, which is an example of a pre-trained artificial intelligence means 131, and the desired result information is displayed on the upload screen 112 of the browser 111, or is output as audio. As a result, the user can obtain the desired information on the result of generating an explanation about the problem in the contents of the specified document file DF or URL information DF1 without reading the content page of the specified document file DF or URL information DF1.

[0048] Incidentally, the prompt buttons on the upload screen 112 of the browser 111 may be configured to have multiple levels of character limit such as a "Summarize! (300 characters)" button, a "Summarize! (500 characters)" button, a "Tell me what the problems are! (300 characters)" button, and a "Tell me what the problems are! (500 characters)" button, so that the instruction command information PT corresponding to each prompt button includes character number condition information for the result information (generated text data GT) to be generated.

[0049] As a result, result information (generated text data GT) with the number of characters desired by the user is displayed on the upload screen 112 of the browser 111, or output as voice. As a result, we can instantly respond to requests, whether they are for detailed information or general information. In particular, when there is no character limit, the pre-trained artificial intelligence means 131 tends to provide a detailed explanation using multiple sentences, so if you want to know the main points or want to reply by capturing the main points, it is recommended that you generate a text with a character limit. By providing buttons with different character limits, the user can obtain result information (generated text data GT) with the number of characters desired in just one go. In other words, in the case of a configuration in which instruction command information PT is sent without a character limit, it is possible to eliminate the trouble of first obtaining result information (generated text data GT) with a relatively large number of characters, and then changing it to result information (generated text data GT) with a relatively small number of characters by limiting the number of characters.

[0050] Furthermore, in this embodiment, when the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data based on the instruction command information PT, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates basis information PF regarding which page or part of the content information identified by the loaded document file DF or URL information DF1 the text data is based on. As the basis information PF, for example, information such as "P15: lines 4 to 10" is generated. Then, the second server 130, which is an example of a server, transmits the generated text data GT and the basis information PF to the user terminal 110 via the first server 120. Then, as shown in FIG. 5, the user terminal 110 is configured to display the generated text data GT and the basis information PF on an upload screen 112 of a browser 111.

[0051] As a result, not only the generated text data GT but also the basis information PF are displayed on the upload screen 112 of the browser 111. As a result, the user can easily check the accuracy and reliability (veracity) of the generated text data GT by referring to the content information specified by the original document file DF or the URL information DF1 based on the evidence information PF. In other words, the user can easily confirm which part and what the generated text data GT was generated based on, without simply looking at the generated text data GT and accepting the contents of the generated text data GT.

[0052] In addition, when the basis information PF displayed on the upload screen 112 is clicked on, the browser 111 may be configured to display the content information identified by the original document file DF or the URL information DF1, and to highlight the basis information PF. This allows users to simply click to highlight the area they want to check. As a result, the user can easily check the portion of the original document file DF or the content information specified by the URL information DF1 that he or she wishes to check.

[0053] In addition, in this embodiment, when the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data based on the instruction command information PT, it generates multiple question information in which information regarding the content information identified by the loaded document file DF or URL information DF1 serves as answer information. For example, as examples of question information, "What if something goes wrong?", "How do you have fun online?", and "How do you have fun in a group?" are generated. Then, the second server 130, which is an example of a server, is configured to transmit the generated text data GT and the generated question information QT to the user terminal 110 via the first server 120 in this order.

[0054] As a result, as shown in Figure 6, multiple anticipated question information about the content information identified by the loaded document file DF or URL information DF1 is generated, and the generated question information QT is displayed in the upload screen 112 of the browser 111 in the order following the generated text data GT. As a result, the user can first select one of the multiple generated question information QT and interact with the pre-trained artificial intelligence means 131 to easily deepen his / her understanding of the content information identified by the document file DF or URL information DF1 read by the pre-trained artificial intelligence means 131 without having to read all of it himself / herself. Here, when an operation to select the generated question information QT is performed, the generated question information QT is sent as a prompt, that is, instruction command information PT, via the first server 120 to the pre-trained artificial intelligence means 131 of the second server 130, and generated text data GT can be obtained as an answer.

[0055] As shown in FIG. 7, in the document file specification field of the upload screen 112 of the browser 111, a plurality of document files DF and URL information DF1 may be specified. The URL information DF1 may be site map information. When a plurality of document files DF and URL information DF1 are specified, all of them are uploaded to the second server 130 via the first server 120. That is, on upload screen 112 of browser 111 of user terminal 110, a plurality of document files DF or a plurality of pieces of URL information DF1 can be specified. If a plurality of document files DF or URL information DF1 are specified, the user terminal 110 uploads the plurality of document files DF or URL information DF1 to the first server 120. As a result, the contents of multiple document files DF or URL information DF1 are read into the chat GPT, which is an example of the pre-trained artificial intelligence means 131. As a result, it is possible to obtain result information on common contents among the contents of a plurality of document files DF or the contents of the URL information DF1, or on the contents of at least one of them. In other words, it is possible to obtain efficient result information that takes into consideration commonalities and differences in the contents of a plurality of document files DF or URL information DF1. When the generated text data GT is displayed on the upload screen 112 as a response to the chat GPT, it is clearly indicated, as basis information PF, which document file DF or which part of the URL information DF1 the information pertains to. That is, even if the user asks a question about the contents of a plurality of document files DF, the basis information PF about the contents of the generated text data GT as the answer clearly indicates which document file DF it is.

[0056] Furthermore, in this embodiment, as shown in Figure 7, if the instruction command information PT is a question about content information identified by the uploaded document file DF or URL information DF1, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, calculates a correlation score between the content of the question and the content of the information regarding the content information identified by the loaded document file DF or URL information DF1. If the calculated correlation score is less than a predetermined threshold, the second server 130 is configured to transmit generated text data GT indicating that the answer is "unknown" to the user terminal 110 via the first server 120 as an answer to the question.

[0057] As a result, when a question is asked about content that has a low correlation with the content information specified by the uploaded document file DF or URL information DF1, an answer that honestly says "I don't know" is displayed on the upload screen 112 of the browser 111, without giving an answer with a low reliability. As a result, the user can understand that the question he or she asked was irrelevant in relation to the uploaded document file DF or the content information specified by the URL information DF1, and can try a different question. In other words, content with a low reliability is not displayed as an answer on the upload screen 112 of the browser 111. As a result, users can interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, with peace of mind.

[0058] In this embodiment, as shown in FIG. 8, when the instruction command information PT includes language designation information PT1 for the answer, for example, assume that the language is designated as "in English!". In this case, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data in the specified language based on the instruction command information PT. The second server 130 is configured to transmit the generated text data GT to the user terminal 110 via the first server 120. As a result, generated text data GT in the specified language is displayed as a response on upload screen 112 of browser 111. As a result, it is possible to accommodate multilingual users.

[0059] Furthermore, in this embodiment, as shown in FIG. 9, when the specified operation is an input operation of an instruction command PT2 in a foreign language other than Japanese, for example, assume that an instruction command information PT to the effect of "Tell me the recipe for Chinese squid!" is input as an instruction command PT2 in English. In this case, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data in English as an example of the language of the input instruction command PT2 based on the instruction command PT2 in English, which is the instruction command information PT. The second server 130 is configured to transmit the generated text data GT to the user terminal 110 via the first server 120. As a result, when an instruction command PT2 such as a question is input in English or another language, text data is generated in that language and the generated text data GT is displayed on the upload screen 112 of the browser 111 as a response. As a result, it is possible to accommodate multilingual users.

[0060] In this embodiment, as shown in FIG. 10, a user terminal 110 communicates with a second server 130 having a chat GPT, which is an example of a pre-trained artificial intelligence means 131, via a first server 120 having a bot 121. As shown in FIG. 11(A), on an administration screen 113 displayed on a browser 111 of a user terminal 110, which is an example of an administrator terminal capable of communicating with the first server 120, masking of word attributes contained in data DT sent from the user terminal 110 to a bot 121 of the first server 120 can be freely set. For example, suppose that the word attributes "name," "address," "telephone number," "My Number," and "email address" are selected and masking is set.

[0061] Then, as shown in FIG. 11(B), assume that an employee list file is specified and input as a document file DF containing characters WD of a masking word attribute that has been preset as an example of data DT on upload screen 112 of browser 111 of user terminal 110, and uploaded to first server 120. Here, an employee list file serving as a document file DF, which is an example of data DT, contains, as employee names, characters with the attribute "name", which is an example of characters WD with a word attribute for masking. It is also assumed that a prompt "Tell me who among these people hasn't been transferred in the last three years" is input as instruction and command information PT and transmitted to the first server 120. Here, the instruction command information PT, which is an example of data DT, does not include any object to be masked. Since the other file, an employee list, contains the masking target, the masking process will be described using the employee list file as an example. As shown in FIG. 12, an employee list file is transmitted from the user terminal 110 to the first server 120.

[0062] Then, the bot 121 of the first server 120 that receives the data determines that the employee list file of the received data DT contains characters WD with the masking word attribute. Here, the bot 121 is configured to be able to determine the attributes of “name,” “address,” “telephone number,” “My Number,” and “email address” when scanning data DT such as text based on the database of the first server 120. Regarding the determination of word attributes, the bot 121 of the first server 120 may refer to a list of word attributes previously stored in a database and determine, for example, that it is a name attribute, an address attribute, etc., or may determine that it is a name attribute, an address attribute, etc. from the context of that part of the text data, or may determine that it is a name attribute, an address attribute, etc. from the way that part of the text data is written (style format).

[0063] More specifically, the names of employees, which are the word attribute of "name" listed in the employee directory, are detected. Then, the bot 121 of the first server 120 converts the employee name, which is the character WD of the masking word attribute, in the employee list file of the received data DT into a combination of "name", which is an example of another arbitrary character WD1 of the same word attribute, and a number such as "1". Thereafter, the bot 121 of the first server 120 transmits the employee list file of the converted data DT1 to the second server 130 together with a prompt. The bot 121 of the first server 120 stores a list of the conversions (a table showing the correspondence between before and after the conversion).

[0064] Thereafter, the second server 130 uses the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data based on the employee list file and prompts of the converted data DT1. As an example, we will select people who have not been transferred for three years for the word attributes "Name 1," "Name 2," etc., which are converted from the joining date information and past transfer history information written in the employee register file. Then, "There are 10 people: Name 1, Name 3, Name 5, Name 7, . . . " is obtained as an example of generated text data GT. Next, the second server 130 transmits the generated text data GT obtained by the execution to the bot 121 of the first server 120.

[0065] Then, the bot 121 of the first server 120 changes (reverse converts) the part of the received generated text data GT that was converted from another arbitrary character WD1 when it was sent to the second server 130 back to the character WD before conversion. The bot 121 of the first server 120 refers to the list (pre-conversion correspondence table) saved at the time of conversion and changes it back to the characters WD before conversion. For example, if "Suzuki Emi" is converted to "Name 1", then "Name 1" is changed back to "Suzuki Emi" (reverse conversion). Then, as an example of the generated text data GT1 after the change, "The ten people are Emi Suzuki, Ichiro Tanaka, John Smith, Amy Marshall, . . . " is obtained. Thereafter, the bot 121 of the first server 120 is configured to transmit the modified generated text data GT1 to the user terminal 110.

[0066] As a result, for example, by setting in advance masking of word attributes that can identify individuals, such as name, address, telephone number, and email address, for data DT sent from the user terminal 110 to the bot 121 of the first server 120, if the data DT entered by the user contains characters WD of those word attributes, the bot 121 of the first server 120 will convert the characters WD of those word attributes into other arbitrary characters WD1 of the same word attribute, and then send the converted data DT1 to the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, and text data generation will be performed based on the converted data DT1. As a result, as shown in FIG. 12, it is possible to prevent the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, from memorizing information about characters WD with word attributes that can identify an individual. In other words, the content sent to the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, is converted into characters that cannot identify individuals. As a result, it is possible to prevent leakage of personally identifiable information to other terminals via the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130.

[0067] Furthermore, when the bot 121 of the first server 120 receives generated text data GT from chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, the part of the received generated text data GT that was converted into another arbitrary character WD1 when sent to chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, is modified to return it to the character WD before conversion, and the modified generated text data GT1 is sent to the user terminal 110. As a result, as shown in FIG. 11(B) and FIG. 12, it is possible to obtain generated text data GT1 after modification that basically conforms to the intent of the command information PT input by the user. In other words, the character WD1 converted when the bot 121 of the first server 120 sends it to the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, is converted back to the original character WD when the bot 121 of the first server 120 receives it from the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130. As a result, it is possible to avoid unintended information leakage to chat GPT, which is an example of pre-trained artificial intelligence means 131 of the second server 130, and unintended information leakage to other terminals via chat GPT, which is an example of pre-trained artificial intelligence means 131 of the second server 130, while ensuring consistency between the instruction and command content from the user terminal 110 and the response content to the user terminal 110.

[0068] Note that, as an example of data DT, an example of masking text data contained in an employee directory file, which is a file for a spreadsheet software with a cell-format display, has been described. However, the data DT to be masked is not limited to files for spreadsheet software. For example, when the data DT is image data such as a photograph or video, the bot 121 uses an image recognition means to determine whether or not there is a masking target, and when it determines that there is a masking target, it masks the target part and replaces it with another part having the same attributes before transmitting it to the second server 130.

[0069] The bot 121 of the first server 120 saves a list of the conversions (a table showing the correspondence between before and after the conversion). Then, the bot 121 changes (reverse-converts) the generated data received from the second server 130 by referring to the list (pre-conversion correspondence table) saved at the time of conversion to return it to its pre-conversion state, and then transmits the changed generated data to the user terminal 110. The generated data and the changed generated data are not limited to text data, but may be image data such as photographs and videos.

[0070] The artificial intelligence system 100, which is an embodiment of the present invention obtained in this manner, comprises a user terminal 110 and a first server 120 and a second server 130 as servers. When an operation to specify a document file DF or an operation to specify URL information DF1, which is location information for predetermined information, is performed on an upload screen 112 of a browser 111 of the user terminal 110, the user terminal 110 uploads the document file DF or the URL information DF1 to the second server 130 via the first server 120. When a predetermined operation is performed on the upload screen 112 of the browser 111, the user terminal 110 transmits instruction command information PT corresponding to the predetermined operation to the second server 130 via the first server 120. The second server 130 then uploads the document file DF or the URL information DF1 to the second server 130 via the first server 120 based on the received instruction command information PT. The system is configured to generate text data for content information specified by file DF or URL information DF1, and transmit the generated text data GT obtained through the generation to user terminal 110 via first server 120, and the user terminal 110 displays the received generated text data GT on upload screen 112 of browser 111 or outputs it as audio.This allows the user to obtain desired generated text data GT information based on prompts of specified command information, such as creating a summary, pointing out problems, pointing out areas for improvement, and questioning information, by simply uploading a document file DF or URL information DF1 and performing a specified operation, without having to read the uploaded document file DF himself or herself, or without having to read the webmail page of the uploaded URL information DF1 or the document file DF specified by the URL information DF1 himself or herself.

[0071] Furthermore, when chat GPT, which is an example of pre-trained artificial intelligence means 131 of the second server 130, generates text data based on instruction command information PT, chat GPT, which is an example of pre-trained artificial intelligence means 131 of the second server 130, generates basis information PF regarding which page or part of the content information identified by the loaded document file DF or URL information DF1 the text data is based on, and the second server 130 transmits the generated text data GT and basis information PF to the user terminal 110 via the first server 120, and the user terminal 110 displays the generated text data GT and basis information PF on the upload screen 112 of the browser 111.This allows the user to easily check the accuracy and reliability (veracity) of the generated text data GT by referring to the original document file DF or the content information identified by the URL information DF1 based on the basis information PF, and the user can easily check which part and what the text data was generated based on, without looking at only the generated text data GT and blindly accepting the contents of the generated text data GT.

[0072] In addition, when the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data based on the instruction command information PT, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates multiple related question information in which information regarding the content information identified by the loaded document file DF or URL information DF1 serves as answer information, and the second server 130 transmits the generated text data GT and the generated question information QT in that order to the user terminal 110 via the first server 120.This allows the user to first select one of the multiple question information and interact with the pre-trained artificial intelligence means 131 to easily deepen his or her understanding of the content information identified by the document file DF or URL information DF1 loaded by the pre-trained artificial intelligence means 131 without having to read all of it himself or herself.

[0073] Furthermore, the instruction command information PT is a question about the content information identified by the uploaded document file DF or URL information DF1, and the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, calculates a correlation score between the content of the question and the content of the information about the content information identified by the loaded document file DF or URL information DF1, and if the calculated correlation score is less than a predetermined threshold, the second server 130 is configured to send generated text data GT to the user terminal 110 via the first server 120, indicating that the answer to the question is unknown.This allows the user to understand that the question he or she asked was off-topic in relation to the content information identified by the uploaded document file DF or URL information DF1 and to try a different question, allowing the user to interact with the pre-trained artificial intelligence means 131 with confidence.

[0074] In addition, when the instruction command information PT includes language designation information PT1 for the answer, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data in the designated language based on the instruction command information PT, and the second server 130 transmits the generated text data GT to the user terminal 110 via the first server 120. When the specified operation is an input operation of an instruction command PT2 in a foreign language other than Japanese, the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, generates text data in the language of the input instruction command PT2 based on the instruction command PT2 in the foreign language, which is the instruction command information PT, and the second server 130 transmits the generated text data GT to the user terminal 110 via the first server 120. This configuration makes it possible to accommodate multilingual users.

[0075] Furthermore, the user terminal 110 communicates with the second server 130 having a chat GPT, which is an example of a pre-trained artificial intelligence means 131, via the first server 120 having the bot 121. In the management screen 113 of the administrator terminal capable of communicating with the first server 120, masking of the word attribute included in the data DT from the user terminal 110 to the bot 121 of the first server 120 is freely set. When the data DT including the character WD of the word attribute of the masking set in advance is inputted in the upload screen 112 of the browser 111 of the user terminal 110 and uploaded to the first server 120, the bot 121 of the first server 120 that receives the data determines that the received data DT includes the character WD of the word attribute of the masking, converts the character WD of the word attribute of the masking in the received data DT into another arbitrary character WD1 of the same word attribute, and transmits the converted data DT1 to the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130. The second server 130 transmits the pre-trained The chat GPT, which is an example of the pre-trained artificial intelligence means 131, is used to generate text data based on the converted data DT1, and the generated text data GT obtained by the execution is transmitted to the bot 121 of the first server 120. The bot 121 of the first server 120 modifies the part of the received generated text data GT, which is another arbitrary character WD1 that was converted when transmitting to the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, back to the character WD before conversion, and then transmits the modified generated text data GT1 to the user terminal 110. This configuration makes it possible to prevent the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, from remembering information about the character WD of a word attribute that can identify an individual, and to prevent leakage of personally identifiable information to other terminals via the chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, and further makes it possible to obtain modified generated text data GT1 that is basically in line with the intention of the instruction command information PT input by the user.It is possible to avoid unintended information leakage to the pre-trained artificial intelligence means 131 of the second server 130 and unintended information leakage to other terminals via the pre-trained artificial intelligence means 131 of the second server 130, and to ensure consistency between the instruction and command contents from the user terminal 110 and the response contents to the user terminal 110.

[0076] Furthermore, the program of the artificial intelligence system 100 which is an embodiment of the present invention includes a file designation determination step S1 for determining whether or not an operation for designating a document file DF or an operation for designating URL information DF1 which is location information for predetermined information has been performed on the upload screen 112 of the browser 111 of the user terminal 110; an upload step S2 for the user terminal 110 uploading the document file DF or the URL information DF1 to the second server 130 via the first server 120 when an operation for designating the document file DF or an operation for designating the URL information DF1 which is location information for the predetermined information has been performed; a predetermined operation determination step S3 for determining whether or not a predetermined operation has been performed on the upload screen 112 of the browser 111; an instruction command transmission step S4 for the user terminal 110 transmitting instruction command information PT according to the predetermined operation to the second server 130 via the first server 120 when the predetermined operation has been performed; The system has a text data generation execution step S5 in which the second server 130 generates text data about the content information specified by the loaded document file DF or URL information DF1, a generated data transmission step S6 in which the second server 130 transmits the generated text data GT obtained by the execution to the user terminal 110 via the first server 120, and a display / audio output step S7 in which the user terminal 110 displays the received generated text data GT on the upload screen 112 of the browser 111 or outputs it as audio. This has a significant effect, as it allows the user to obtain desired information about the generated text data GT based on prompts from specified command information, such as making a summary, pointing out problems, pointing out areas for improvement, and asking questions about information, simply by uploading the document file DF or URL information DF1 and performing a specified operation, without having to read the uploaded document file DF himself or herself, or without having to read the webmail page of the uploaded URL information DF1 or the document file DF specified by the URL information DF1 himself or herself. [Explanation of symbols]

[0077] 100 ··· Artificial Intelligence Systems 110 User terminal 111 ··· Browser 112 Upload screen 113: Administration screen (of administrator's terminal) 120 First Server (Server) 121 ··· Bot 130 Second Server (Server) 131 ··· Pre-trained AI method (Chat GPT) DF Document file DF1 ··· URL information PT: Prompt PT1 Language specification information for the answer PT2: Commands and instructions in a foreign language GT: Generated text data GT1: Generated text data after modification PF: Basis information QT ··· Generate question information DT...Data DT1: Post-conversion data (post-conversion instruction command information) WD: (predefined) masking word attribute character WD1: Any other character with the same word attribute

Claims

1. An artificial intelligence system comprising a user terminal and a server, the user terminal displaying or outputting audio based on the server, When an operation for designating a document file or an operation for designating URL information, which is location information for predetermined information, is performed in a browser of the user terminal, the user terminal uploads the document file or the URL information to a server; When a predetermined operation is performed on the browser, the user terminal transmits instruction command information corresponding to the predetermined operation to a server; The server uses the pre-trained artificial intelligence means based on the received instruction command information to generate text data for the content information specified by the loaded document file or URL information, and transmits the generated text data to the user terminal; An artificial intelligence system characterized in that the user terminal is configured to display the received generated text data on a browser or output it as voice.

2. When the pre-trained artificial intelligence means generates text data based on instruction command information, the pre-trained artificial intelligence means generates basis information regarding which page or part of the content information specified by the loaded document file or URL information the text data is based on; The server transmits the generated text data and the grounds information to a user terminal; 2. The artificial intelligence system according to claim 1, wherein the user terminal is configured to display the generated text data and the basis information on a browser.

3. When the pre-trained artificial intelligence means generates text data based on instruction command information, the pre-trained artificial intelligence means generates a plurality of pieces of question information in which information on content information specified by the loaded document file or URL information serves as answer information; 3. The artificial intelligence system according to claim 2, wherein the server is configured to transmit the generated text data and the generated question information to the user terminal in that order.

4. The instruction and command information is a question about an uploaded document file or content information identified by URL information, The pre-trained artificial intelligence means calculates a correlation score between the content of the question and the content of the information regarding the content information specified by the loaded document file or URL information, The artificial intelligence system described in claim 3, characterized in that if the calculated correlation score is less than a predetermined threshold, the server is configured to send generated text data to the user terminal stating that the answer to the question is unknown.

5. When the instruction command information includes language designation information for a response, the pre-trained artificial intelligence means generates text data in the designated language based on the instruction command information; The server transmits the generated text data to a user terminal; When the predetermined operation is an input operation of an instruction command in a foreign language other than Japanese, the pre-trained artificial intelligence means generates text data in the language of the input instruction command based on the instruction command information, 5. The artificial intelligence system according to claim 4, wherein the server is configured to transmit the generated text data to a user terminal.

6. The user terminal communicates with a second server having a pre-trained artificial intelligence means via a first server having a bot; In a management screen of an administrator terminal capable of communicating with the first server, masking of word attributes included in data from the user terminal to the bot of the first server can be freely set, When data including characters of a masking word attribute preset in the browser of the user terminal is input and uploaded to the first server, the bot of the first server that receives the data determines that the received data includes characters of a masking word attribute, converts the characters of the masking word attribute in the received data into other arbitrary characters of the same word attribute, and then transmits the converted data to the pre-trained artificial intelligence means of the second server; The second server uses the pre-trained artificial intelligence means of the second server to generate text data based on the converted data, and transmits the generated text data to the bot of the first server; An artificial intelligence system as described in any one of claims 1 to 5, characterized in that the bot of the first server is configured to change any other arbitrary character part of the received generated text data that was converted when sending it to the pre-trained artificial intelligence means of the second server back to the character before conversion, and then send the changed generated text data to the user terminal.

7. The instruction information is to summarize, The artificial intelligence system described in claim 6, characterized in that the second server uses a pre-trained artificial intelligence means to generate text data for the content information identified by the loaded document file or URL information based on the received instruction command information to summarize, and transmits the generated text data obtained by the execution to a user terminal via the first server.

8. The instruction and command information checks whether there is a problem and, if there is a problem, explains the content of the problem, The artificial intelligence system described in claim 6, characterized in that the second server is configured to use a pre-trained artificial intelligence means to generate text data for the content information identified by the loaded document file or URL information based on the received instruction command information to check whether there are any problems and, if there are any problems, to explain the content of the problems, and to transmit the generated text data obtained by the execution to the user terminal via the first server.

9. The artificial intelligence system described in claim 6, characterized in that when the number of document files or URL information that can be specified in the browser of the user terminal is multiple, and multiple document files or URL information are specified, the user terminal is configured to upload the multiple document files or URL information to the first server.

10. A program of an artificial intelligence system that a user terminal displays or outputs voice based on a server, a file designation determination step of determining whether or not an operation for designating a document file or an operation for designating URL information, which is location information for predetermined information, has been performed in a browser of the user terminal; an uploading step in which, when an operation for designating the document file or an operation for designating URL information, which is location information for the predetermined information, is performed, the user terminal uploads the document file or the URL information to the server; a predetermined operation determination step of determining whether or not a predetermined operation has been performed on the browser; an instruction command sending step of, when the predetermined operation is performed, sending instruction command information corresponding to the predetermined operation to a server by the user terminal; a text data generation execution step in which the server executes text data generation for the content information specified by the loaded document file or URL information by using a pre-trained artificial intelligence means based on the received instruction command information; a generated data transmission step in which the server transmits generated text data obtained by the execution to a user terminal; A program for an artificial intelligence system, comprising a display / audio output step in which the user terminal displays the received generated text data on a browser or outputs it as audio.

Images