Information processing system, terminal device, information processing method, and program

Abstract

To more easily obtain a work result with high accuracy using artificial intelligence.SOLUTION: In a user terminal 10 of an information processing system 1, an UI display control section 151 receives a work request to a medication history management server 20 by natural language from a user. A prompt generation section 154 transmits contents of the received work request, interface information (API) on information processing which can be used in the medication history management server 20, and a sample of control information to the medication history management server 20 using the interface information to a generation type AI service providing server 40 as a prompt. A medication history management system control section 155 acquires answer information transmitted from the generation type AI service providing server 40 according to transmission of the prompt, and transmits it as the control information to the medication history management server 20. The medication history management server 20 performs work according to the transmitted control information, and transmits a work result to the user terminal 10.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to an information processing system, a terminal device, an information processing method, and a program. [Background technology]

[0002] In recent years, there has been a surge in efforts to use artificial intelligence to assist with tasks performed by humans or to have artificial intelligence perform tasks on behalf of humans. In such technology, the procedures for performing a task are incorporated into artificial intelligence as rules, and the actual task is then executed according to these rules; or a machine learning model is prepared that has been trained on sample tasks that serve as training data, and the actual tasks are then input into the machine learning model to infer the results of the tasks. Patent Document 1 describes a technology that uses artificial intelligence to assist with work. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-064183 Summary of the Invention [Problem to be solved by the invention]

[0004] However, when incorporating the procedure for performing a task as rules into an artificial intelligence, there are problems in that it is cumbersome to accurately incorporate the rules into the artificial intelligence, and that the method lacks applicability. In addition, when inferring work results using a machine learning model, the inferred work results are not necessarily accurate, and it is not easy to reflect specific conditions in the inference, etc. In particular, in fields related to medicine, when obtaining work results using artificial intelligence, higher accuracy is required, and it may not be appropriate to immediately apply technology using existing machine learning models. As described above, in conventional techniques, it has been difficult to easily obtain highly accurate work results using artificial intelligence.

[0005] The objective of the present invention is to utilize artificial intelligence to more easily obtain highly accurate work results. [Means for solving the problem]

[0006] In order to solve the above problems, an information processing system according to one aspect of the present invention comprises: An information processing system including a terminal device used by a user and a server that executes an operation in response to a request from the terminal device, The terminal device a request receiving means for receiving a task request from a user to the server in natural language; an answer request means for transmitting the contents of the work request received by the request receiving means, interface information of information processing available on the server, and a sample of control information for the server using the interface information as a prompt to the generative AI; A task request sending means for acquiring answer information sent from the generative AI in response to sending the prompt, and sending the answer information as control information to the server; Equipped with The server, The work is carried out in accordance with the control information transmitted from the work request transmission means, and the work results are transmitted to the terminal device. Effect of the Invention

[0007] According to the present invention, by utilizing artificial intelligence, highly accurate work results can be obtained more easily. [Brief description of the drawings]

[0008] [Figure 1] 1 is a diagram showing a system configuration of an information processing system 1 according to an embodiment of the present invention. [Diagram 2]1 is a schematic diagram showing an example of a specific format of a prompt sent from a user terminal 10 to a generation-type AI service providing server 40. FIG. [Diagram 3] FIG. 2 is a schematic diagram showing a specific example of an execution plan (control information). [Figure 4] 2 is a schematic diagram showing an example of an output screen displayed on the user terminal 10. FIG. [Diagram 5] FIG. 8 is a diagram showing a hardware configuration of an information processing device 800 constituting each device. [Figure 6] FIG. 2 is a block diagram showing the functional configuration of a user terminal 10. [Figure 7] 1 is a flowchart showing the flow of a medication history management system control process executed by the information processing system 1. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0009] Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

[0010] The information processing system of this embodiment uses generative AI combined with a large-scale language model (LLM) to generate control information for controlling a specific application system (here, a medical history management system) in response to a user's voice or text (i.e., natural language) request. At this time, text data expressing the user's request, a set of API (Application Programming Interface) data available in a specific application system, and sample control information for controlling a specific application via the API are input as prompts for the generative AI. This makes it possible to generate control information (control information for controlling a specific application system) that can obtain work results in response to a user request expressed in natural language by inputting prompts written in accordance with specified conditions to a generative AI with general functions, without having the generative AI learn information regarding the control of a specific application system. Therefore, according to the present invention, it is possible to more easily obtain highly accurate work results by utilizing artificial intelligence. A specific configuration will be described below.

[0011] [System configuration of information processing system 1] FIG. 1 is a diagram showing a system configuration of an information processing system 1 according to an embodiment of the present invention. As shown in FIG. 1, the information processing system 1 includes a user terminal 10, a medication history management server 20, a voice recognition server 30, and a generative AI service providing server 40, and the user terminal 10, the medication history management server 20, the voice recognition server 30, and the generative AI service providing server 40 are configured to be able to communicate with each other via a network 50. In the information processing system 1 according to this embodiment, the user terminal 10 functions as a client terminal for using the medication history management server 20. In addition, in FIG. 1, the system configuration of the information processing system 1 is shown as including the voice recognition server 30 and the generative AI service providing server 40, but these servers can also use servers that provide general-purpose services operated by a third party other than the operator of the medication history management server 20. In addition, each server can be configured by one server computer, or can be configured as a virtual server in a server system configured by multiple server computers.

[0012] The user terminal 10 is a terminal device used by a user who utilizes the medical history management system constituted by the medical history management server 20, and is constituted by an information processing device such as a smartphone, a mobile phone, a tablet terminal, or a PC (Personal Computer). The user terminal 10 is capable of performing operations to utilize the functions of the medical history management server 20 via input devices such as a mouse and a keyboard, and is also capable of inputting the work requested by the user in natural language by voice input (or text input). Hereinafter, information representing the work requested by the user to the medical history management server 20, input as natural language by voice (or text), is referred to as an "operation instruction sentence".

[0013] When an operation instruction statement is input by voice input, the user terminal 10 transmits voice data to the voice recognition server 30, and receives character string data representing the operation instruction statement obtained by voice recognition processing.

[0014] In addition, the user terminal 10 has previously acquired APIs (interface information for information processing) that can be used in the medical history management server 20, and stores data of sample control information (control codes representing specified operations expressed using the API) for causing the medical history management server 20 to perform specified tasks using the API. Then, the user terminal 10 sends a prompt to the generative AI service providing server 40, which includes an instruction to generate control information for causing the medical history management server 20 to execute the task represented by the operation instruction statement, text data representing the operation instruction statement input by voice (or text) (for example, "Enter the concomitant medication □□ into Mr. Xyama △o's medical history"), a set of API data available on the medical history management server 20, and sample control information (control code representing a specified operation expressed using the API) for controlling the medical history management server 20 via the API. Furthermore, the user terminal 10 receives answer information (control information for the medical history management server 20) transmitted from the generation type AI service providing server 40 in response to transmitting the prompt.

[0015] In addition, the user terminal 10 transmits the answer information (control information) received from the generation type AI service providing server 40 to the medical history management server 20. The answer information (i.e., control information for the medication history management server 20) transmitted from the user terminal 10 corresponds to the control information when the user operates the medication history management server 20 via an input device. Therefore, even when answer information (control information) is transmitted, the medication history management server 20 can execute processing in the same manner as when the user operates the medication history management server 20 via an input device, and return the processing result to the user terminal 10.

[0016] Fig. 2 is a schematic diagram showing an example of a specific format of a prompt sent from the user terminal 10 to the generation-type AI service providing server 40. In Fig. 2, the control information of the medication history management server 20 generated by the prompt is referred to as an "execution plan." As shown in FIG. 2, the prompt sent to the generative AI service providing server 40 includes the following: (1) "Instruction to create an execution plan" which indicates an instruction to generate control information for causing the medical history management server 20 to execute the task indicated by the operation instruction statement (2) "Definition of functions (interfaces) that can be used in the execution plan" representing a set of API data that can be used by the medical history management server 20 (3) "Execution plan creation rules" that represent the conditions for generating control information (4) "Execution plan output result format specification and rules" for specifying the output format of control information (5) "Natural language instructions from the user" that express the work content specified in the operation instruction sentence Each item is defined. By sending a prompt including these items to the generative AI service providing server 40, the generative AI service providing server 40 uses the API defined in the prompt to generate control information (execution plan) for the medication history management server 20, which indicates the work content specified in the operation instruction statement, according to the specified rules. When this control information (execution plan) is executed by the medication history management server 20, the processing result specified in the operation instruction statement is returned.

[0017] The medical history management server 20 is configured by a PC or a server computer, and functions as a medical history management system that performs medical history management work in a pharmacy. Note that the medical history management system is an example of a business application, and the medical history management server 20 may be provided with functions for performing various tasks other than the medical history management task (various business applications such as a prescription computer or an electronic medical record system) as appropriate. For example, the medical history management server 20 may have a medical history management function for managing the contents of prescriptions issued to patients and the history of drugs provided to patients, or a medical fee statement (receipt) creation function for calculating medical fees. The functions of the medical history management server 20 may be further expanded to include an electronic medical record management function for managing electronic medical records, and a test result database function for managing test result data such as patient test values ​​and CT (Computed Tomography) images. As described above, the medical history management server 20 can be operated from the user terminal 10 using an API, and can accept requests in the same way and return processing results based on the medical history management function, etc., even when a request is made by the API, when an operation is made from the user terminal 10 via an input device, or when answer information generated by the generative AI service providing server 40 is sent and an operation is requested.

[0018] The voice recognition server 30 is configured by a server computer, receives voice data transmitted from the user terminal 10, and executes voice recognition processing based on natural language processing on the received voice data. Then, the voice recognition server 30 transmits to the user terminal 10 a character string (a sentence expressed in natural language) corresponding to the voice data acquired by the voice recognition processing.

[0019] The generative AI service providing server 40 is configured by a server computer and implements generative AI using a large-scale language model (LLM). In response to a prompt (information representing a request to AI) received from the user terminal 10, the generative AI service providing server 40 refers to the large-scale language model and generates information that is a response to the prompt. In this embodiment, the prompt transmitted from the user terminal 10 to the generative AI service providing server 40 includes text data representing an operation instruction sentence (for example, "Enter concomitant medication □□ in Mr. XX's medical history") for the medical history management server 20, a data group of APIs available in the medical history management server 20, and sample control information (control code representing a predetermined operation expressed using the API) for controlling the medical history management server 20 by the API (see FIG. 2). In this embodiment, the generative AI service providing server 40 that receives the prompt uses the API defined in the prompt in accordance with the contents of the prompt and generates control information (execution plan) for the medical history management server 20 representing the work content specified in the operation instruction statement according to specified rules. Then, the generation-type AI service providing server 40 transmits the answer information generated in response to the prompt to the user terminal 10.

[0020] [Example of prompt] In this embodiment, the prompt sent to the generative AI service providing server 40 includes the following items, as shown in the format example in Figure 2: "Instructions for creating an execution plan," "Definition of functions (interfaces) that can be used in the execution plan," "Rules for creating an execution plan," "Designation and rules for the execution plan output result format," and "Natural language instructions from the user." Note that the specific description form for describing the instructions that make up the prompt can be any description form that can be interpreted by the generative AI service providing server 40, such as an ordinary language such as English or Japanese, or a specified programming language. Specifically, as an "instruction for creating an execution plan", a directive is written which indicates "Create an execution plan in XML that satisfies the goal conditions step by step. Use available functions."

[0021] In addition, as a "definition of functions (interfaces) that can be used in the execution plan", [Available functions] (1) Data format conversion function {...} (2) Current date and time acquisition function {···} (3) Medication history screen open function {...} Various functions (functions that can be used in the medical history management system) are described.

[0022] In addition, the "rules for creating an execution plan" are as follows: Create an action plan by following these steps: (i) Keep the implementation plan as simple as possible. (ii) Develop an execution plan as a sequence of functions that achieve the goal. (iii)..." The various rules are listed below.

[0023] In addition, as "Specifying and rules for the execution plan output result format", Create your entire implementation plan in the form below. <Plan start> <Function name:...> ... <Plan End> The function should be called by following these steps: (i) The function has one or more named parameters and one output that is all strings. The parameter values ​​are xml-escaped. (ii) Use the ··· function to save the output from a function and pass it to the next function. (iii)..." Specifications and rules such as the above are listed.

[0024] In addition, as "natural language instructions from the user", <goal> Add Bayer Aspirin to the concomitant medications, enter height 175.6cm, weight 62.3kg< / goal> is stated.

[0025] [Example of execution plan (control information)] Next, a specific example of an execution plan (control information) generated in response to a prompt will be described. FIG. 3 is a schematic diagram showing a specific example of an execution plan (control information). As shown in FIG. 3, the execution plan (control information) generated by the generation-type AI service providing server 40 includes the following contents: <plan> <!-- First, we need to add the drug "バイアスピリン" to the list of concurrent drugs. --> <function.***.editInspectionItem server_url="http: / / localhost:**** / **** / api" payload='{"inspectionItems":[{"itemType":"Concurrent drug","itemValue":"Biasprin"}]}' content_type="application / json" / > <!-- Next, we need to input the patient's height. --> <function.***.editInspectionItem server_url="http: / / localhost:**** / **** / api" payload='{"inspectionItems":[{"itemType":"Height","itemValue":"175.6 cm"}]}' content_type="application / json" / > <!-- Finally, we need to input the patient's weight. --> <function.****.editInspectionItem server_url="http: / / localhost:**** / **** / api" payload='{"inspectionItems":[{"itemType":"Weight","itemValue":"62.3 kg"}]}' content_type="application / json" / > Note: In the translation, for the Japanese drug name "バイアスピリン", I used "Biasprin" as a possible equivalent. You may need to verify if there is a more accurate English name for this drug in the relevant medical context. Also, the server URL parts with **** are left as they are as they seem to be specific to a system and not directly translatable without more context.< / plan> That is, in the example shown in Figure 3, an execution plan (control information) is generated that expresses the work content requested in the operation instruction statement using APIs that can be used in the medical history management server 20, and this execution plan is executed in the medical history management server 20 to output the work results requested by the user.

[0026] [Example of output screen] FIG. 4 is a schematic diagram showing an example of an output screen displayed on the user terminal 10. As shown in FIG. As shown in FIG. 4, the output screen (rightmost area) of the medical history management system in this embodiment displays operation instructions in natural language input by the user via voice (or text), as well as answer information (control information) generated in response to the operation instructions. In addition, on the output screen of the medical history management system, the area (other than the rightmost area) where information regarding the patient's medical history is displayed displays the output results returned from the medical history management system based on response information (control information) to the operation instruction statement. Therefore, the user can check the operation instruction statement that he or she inputted, the answer information (control information) generated in response to the operation instruction statement, and the output result returned from the medical history management system based on the control information, and can easily determine whether appropriate content has been output as the intended work result.

[0027] [Hardware configuration] Next, the hardware configuration of each device in the information processing system 1 will be described. In the information processing system 1, each device is configured by an information processing device such as a PC, a server computer, or a tablet terminal, and the basic configurations of the devices are the same.

[0028] FIG. 5 is a diagram showing a hardware configuration of an information processing device 800 constituting each device. As shown in FIG. 5, the information processing device 800 constituting each device includes a CPU (Central Processing Unit) 811, a ROM (Read Only Memory) 812, a RAM (Random Access Memory) 813, a bus 814, an input unit 815, an output unit 816, a memory unit 817, a communication unit 818, a drive 819, and an imaging unit 820.

[0029] The CPU 811 executes various processes according to a program recorded in a ROM 812 or a program loaded from a storage unit 817 to a RAM 813 . The RAM 813 also stores data and the like necessary for the CPU 811 to execute various processes.

[0030] The CPU 811, the ROM 812, and the RAM 813 are connected to one another via a bus 814. To the bus 814, an input unit 815, an output unit 816, a storage unit 817, a communication unit 818, a drive 819, and an imaging unit 820 are connected.

[0031] The input unit 815 is composed of various buttons, a microphone, etc., and is used to input various information in response to instruction operations. The output unit 816 is composed of a display, a speaker, etc., and outputs images and sounds. The storage unit 817 is composed of a hard disk or a dynamic random access memory (DRAM) or the like, and stores various data managed by each server. The communication unit 818 controls communications with other devices via the network 50 .

[0032] Removable media 831, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is appropriately loaded into the drive 819. A program read from the removable media 831 by the drive 819 is installed in the storage unit 817 as necessary. The imaging section 820 is configured with an imaging device including a lens, an imaging element, etc., and captures a digital image of a subject. When the information processing device 800 is configured as the medical history management server 20, the voice recognition server 30, or the generative AI service providing server 40, it is also possible to omit the imaging unit 820. When the information processing device 800 is configured as a tablet terminal, it is also possible to configure the input unit 815 as a touch sensor and place it on top of the display of the output unit 816 to provide a touch panel.

[0033] [Functional configuration] Next, the functional configuration of the information processing system 1 will be described.

[0034] [Functional configuration of user terminal 10] FIG. 6 is a block diagram showing the functional configuration of the user terminal 10. As shown in FIG. 6, a user interface display control unit (UI display control unit) 151, an API set acquisition unit 152, a voice recognition processing unit 153, a prompt generation unit 154, and a medication history management system control unit 155 function in a CPU 811 of the user terminal 10. In addition, an API set storage unit 171, a sample information storage unit 172, a prompt storage unit 173, an answer information storage unit 174, and a task result storage unit 175 are formed in the storage unit 817 of the user terminal 10.

[0035] The API set storage unit 171 stores a data set of APIs that can be used by the medication history management server 20. The data set of APIs stored in the API set storage unit 171 is referenced when the user directly operates the medication history management server 20 with an input device while viewing a UI screen, and the necessary APIs are used. In addition, when the user inputs an operation to be requested of the medication history management server 20 by voice input (or text input), the data set of APIs stored in the API set storage unit 171 is referenced for the purpose of generating a prompt for the generative AI service providing server 40, and some or all of the data set of APIs is included in the prompt and transmitted.

[0036] The sample information storage unit 172 stores various control information (control codes representing various operations represented using the API) serving as samples for controlling the medication history management server 20 by the API. For example, the sample information storage unit 172 stores control information representing an example of an operation for adding a concomitant medication to the medication history of a specific patient, control information representing an example of an operation for adding medication instructions to the medication history of a specific patient, or control information representing an example of an operation for reading and printing data of a medication information document of a specific patient.

[0037] The prompt storage unit 173 stores data on prompts transmitted from the user terminal 10 to the generation-type AI service providing server 40 as history. Data on previously generated prompts can be used appropriately thereafter when a new prompt is generated, for example by referencing a similar prompt.

[0038] The answer information storage unit 174 stores answer information data transmitted from the AI ​​service providing server 40 in response to a prompt as history in association with the transmitted prompt. The work result memory unit 175 stores data on the processing results (work results) of the medical history management server 20 as history when a user operates the medical history management server 20 via an input device and when the user operates the medical history management server 20 using voice (or text) operation instructions.

[0039] The UI display control unit 151 displays a user interface screen (UI screen) generated for the user to execute various processes on the user terminal 10, and displays a screen for performing operations on the medical history management server 20 via an input device, and a screen for inputting a task requested by the user in natural language by voice input (or text input) to operate the medical history management server 20. The UI display control unit 151 also displays a screen showing the processing result received from the medical history management server 20.

[0040] The API set acquisition unit 152 acquires a data group of the API provided by the medication history management server 20, and stores it in the API set storage unit 171. Note that the API set acquisition unit 152 acquires a data group of the API in advance and stores it in the API set storage unit 171 when the user terminal 10 uses the functions of the medication history management server 20.

[0041] When the user of the user terminal 10 inputs an operation to the medication history management server 20 by voice, the voice recognition processing unit 153 records the voice data and transmits the voice data to the voice recognition server 30. In addition, the voice recognition processing unit 153 receives a voice recognition result (text data representing an operation instruction sentence) from the voice recognition server 30 and passes it to the prompt generation unit 154.

[0042] The prompt generation unit 154 adds a set of API data available in the medication history management server 20 and sample control information (control code representing a specific operation expressed using the API) for controlling the medication history management server 20 by the API to the text data representing the operation instruction sentence passed from the voice recognition processing unit 153, and transmits the result as a prompt to the generative AI service providing server 40. Note that the format of the prompt can be in various forms, and for example, as shown in FIG. 2, the format can include "instruction to create an execution plan," "definition of functions (interfaces) that can be used in the execution plan," "rules for creating an execution plan," "designation and rules for the format of the execution plan output result," and "instructions in natural language from the user." In addition, the prompt generation unit 154 receives answer information (control information) transmitted from the generation-type AI service providing server 40 in response to the transmitted prompt, and passes it to the medication history management system control unit 155.

[0043] The medication history management system control unit 155 transmits the answer information (control information) received from the prompt generating unit 154 as a task request to the medication history management server 20. Even when the medication history management system control unit 155 transmits the answer information (control information) to the medication history management server 20, a processing result is returned to the user terminal 10 in the same manner as when the user visually checks the UI screen and operates the medication history management server 20 via an input device such as a mouse.

[0044] [Operation] Next, the operation of the information processing system 1 will be described.

[0045] [Medical history management system control processing] FIG. 7 is a flowchart showing the flow of the medication history management system control process executed by the information processing system 1. The medical history management system control process is started in the user device 10 in response to an instruction to execute the medical history management system control process.

[0046] When the medical history management system control process is started, a data group of APIs provided by the medical history management server 20 is acquired and stored in the API set storage unit 171 in step S1. In step S2, the UI display control unit 151 displays a UI screen generated for the user to execute various processes on the user terminal 10. In step S3, the voice recognition processing unit 153 determines whether or not an operation instruction (voice or text) has been input to the medical history management server 20 by the user. If the user has not input an operation instruction to the medical history management server 20, the result of step S3 is determined to be NO, and the process proceeds to step S4. On the other hand, if the user has input an operation instruction statement to the medical history management server 20, the result of step S3 is determined to be YES, and the process proceeds to step S5.

[0047] In step S4, the medical history management system control unit 155 executes an operation of the medical history management server 20 based on the operation content on the UI screen via the input device. After step S4, the process proceeds to step S11. In step S5, the voice recognition processor 153 determines whether or not an operation instruction has been input by voice. If an operation instruction has not been input by voice, the result of step S5 is NO, and the process proceeds to step S7. On the other hand, if an operation instruction has been input by voice, the result of the determination in step S5 is YES, and the process proceeds to step S6. In step S6, the voice recognition processing unit 153 transmits the voice (voice data of the operation instruction) input by the user for operating the medical history management server 20 to the voice recognition server 30, and executes voice recognition processing. In response to this, text data representing the operation instruction is transmitted from the voice recognition server 30.

[0048] In step S7, the prompt generation unit 154 adds a set of API data available to the medication history management server 20 and sample control information (control code representing a specific operation expressed using the API) for controlling the medication history management server 20 via the API to the text data representing the operation instruction sentence, and generates a prompt for the generative AI service providing server 40. In step S8, the prompt generation unit 154 transmits the generated prompt to the generation AI service providing server 40. In response, answer information (control information) for controlling the medical history management server 20 is transmitted from the generation AI service providing server 40.

[0049] In step S9, the medication history management system control unit 155 acquires the answer information (control information) generated by the generation-type AI service providing server 40. In step S10, the medical history management system control unit 155 transmits answer information (control information) to the medical history management server 20 as a task request. In step S11, the UI display control unit 151 determines whether or not an instruction to end the medical history management system control process has been given. If the end of the medication history management system control process has not been instructed, the result of step S11 is determined to be NO, and the process proceeds to step S2. On the other hand, if an instruction to end the medical history management system control process is given, the result of the determination in step S11 is YES, and the medical history management system control process ends.

[0050] By such processing, in the information processing system 1, when an operation based on natural language is performed on the medication history management server 20 by voice input (or text input), an API available on the medication history management server 20 and sample control information using this API, together with data on the operation instruction text, can be sent to the generative AI service providing server 40, thereby obtaining control information capable of appropriately executing operations on the medication history management server 20 expressed in natural language. Then, the medical history management server 20 is operated based on this control information, and the work results requested by the user can be obtained from the medical history management server 20. In other words, by inputting a prompt written in accordance with specified conditions to a generative AI with general functions, without having the generative AI learn information regarding the control of a specific application system, it is possible to generate control information (control information for controlling a specific application system) that can obtain work results in response to a user request expressed in natural language. Therefore, according to the present invention, it is possible to more easily obtain highly accurate work results by utilizing artificial intelligence.

[0051] [Variation 1] In the above embodiment, an example has been described in which the work content to be input by the user in natural language is the work executed by the medical history management server 20 (medical history management system), but the present invention is not limited to this. For example, if the medication history management server 20 has the function of linking with other systems such as a prescription computer, an electronic medical record system, or a server system responsible for community-based comprehensive medical care, the user can input in natural language to request tasks to be performed by other systems that link with the medication history management server 20. In other words, by using the medical history management server 20 as an interface for making a work request in natural language, a work request can be indirectly made in natural language to a system other than the medical history management server 20. This makes it possible to accept a user's natural language work request without preparing an environment for making work requests in natural language in all systems used by the user.

[0052] [Variation 2] In the above-described embodiment, the task requested by the user in natural language may be a single task executed by the medical history management server 20, or may be a series of tasks that combine a plurality of tasks. In other words, a series of tasks such as "processing to add concomitant medications to the medical history" and "processing to add test value data related to the concomitant medication to the medical history" can be defined as tasks that the user requests in natural language, and a unique name (e.g., "add concomitant medications and test values") can be given to the defined series of tasks. When the user inputs the unique name given to the above series of tasks in natural language into the user terminal 10, the user terminal 10 recognizes the defined series of tasks and sends a prompt to the generative AI service providing server 40 to generate control information for causing the medical history management server 20 to execute this series of tasks. This allows the user to more easily obtain the results of a plurality of tasks that are routinely performed by inputting a combination of a plurality of tasks as a single task request (operation instruction sentence) in natural language.

[0053] As described above, the information processing system 1 according to this embodiment includes the user terminal 10 and the medical history management server 20. The user terminal 10 includes a UI display control unit 151, a prompt generation unit 154, and a medication history management system control unit 155. The UI display control unit 151 accepts an operation request to the medical history management server 20 from the user in natural language. The prompt generation unit 154 transmits the contents of the work request accepted by the UI display control unit 151, interface information (API) for information processing that can be used by the medical history management server 20, and a sample of control information for the medical history management server 20 using the interface information as a prompt to the generative AI service providing server 40. In response to the prompt, the medical history management system control unit 155 acquires the answer information sent from the generation type AI service providing server 40 and sends it as control information to the medical history management server 20. The medical history management server 20 executes operations according to the control information transmitted from the medical history management system control unit 155, and transmits the operation results to the user terminal . As a result, when an operation based on natural language is performed on the medication history management server 20 via voice input (or text input), interface information (API) available on the medication history management server 20 and sample control information using this interface information (API) along with data on the operation instruction text can be sent to the generative AI service providing server 40, thereby obtaining control information capable of appropriately executing operations on the medication history management server 20 expressed in natural language. Then, the medical history management server 20 is operated based on this control information, and the work results requested by the user can be obtained from the medical history management server 20. Therefore, by utilizing artificial intelligence, it becomes possible to more easily obtain highly accurate work results.

[0054] In response to the transmission of the control information, the UI display control unit 151 of the user terminal 10 outputs the work result transmitted from the medication history management server 20, the control information, and a work request (operation instruction text) to the medication history management server in natural language input by the user when the control information was generated. This allows the user to check the operation instruction statement they input, the answer information (control information) generated in response to the operation instruction statement, and the output result returned from the medical history management system based on the control information, and allows the user to easily determine whether appropriate content is being output as the intended work result.

[0055] The prompt generating unit 154 includes information specifying the output format of the answer information in the prompt and transmits it to the generation type AI. The medical history management system control unit 155 acquires the answer information output in that output format from the generative AI service providing server 40, and transmits the answer information to the medical history management server 20 as control information. This makes it possible to easily obtain control information in a format that can appropriately instruct the user to perform the task requested.

[0056] The prompt generation unit 154 includes information specifying the rules for generating answer information in the prompt and transmits it to the generation-type AI service providing server 40. The medical history management system control unit 155 acquires answer information generated according to the rules from the generation type AI service providing server 40, and transmits the answer information to the medical history management server 20 as control information. This allows the generation type AI service providing server 40 to generate control information according to rules assumed by the user.

[0057] As a task expressed in natural language, a task name that represents a series of tasks that combine a plurality of tasks is defined. The UI display control unit 151 accepts a task name as a task request to the medical history management server 20 from the user in natural language. The prompt generation unit 154 transmits prompts for acquiring answer information for each of a series of tasks defined as task names to the generation-type AI service providing server 40. The medication history management system control unit 155 acquires the answer information sent from the generative AI service providing server 40 and transmits it to the medication history management server 20 as control information for acquiring the work results of each series of tasks defined as task names. This allows the user to more easily obtain the results of a plurality of tasks that are routinely performed by inputting a combination of a plurality of tasks as a single task request (operation instruction sentence) in natural language.

[0058] The medical history management server 20 has a function for linking with other systems. When a user sends a work request to the medication history management server 20 in natural language that is related to other systems, the medication history management server 20 performs the work in accordance with the control information, performs the work related to the other systems, and transmits the work results including the work related to the other systems to the user terminal 10. This makes it possible to accept a user's natural language work request without preparing an environment for making work requests in natural language in all systems used by the user.

[0059] The present invention is not limited to the above-described embodiment, and modifications and improvements within the scope of the present invention that can achieve the object of the present invention are included in the present invention. For example, the medical history management server 20, the voice recognition server 30, and the generative AI service providing server 40 in the above-described embodiments can be configured as an on-premise server or a cloud server. In addition, the information processing system 1 according to the present invention may be configured using each device in the above-described embodiments, or the functions of multiple devices in the above-described embodiments may be implemented together in one device, or the functions of one device may be distributed and implemented across multiple devices.

[0060] The above-described series of processes can be executed by hardware or software. In other words, the functional configurations in the above-described embodiments are merely examples and are not particularly limited. That is, it is sufficient that any of the computers constituting the information processing system 1 has a function capable of executing the above-described series of processes as a whole, and the type of functional block used to realize this function is not particularly limited to the examples shown. Furthermore, one functional block may be configured as a single piece of hardware, a single piece of software, or a combination of both.

[0061] In addition, the recording medium containing the program for executing the series of processes described above may be composed not only of a removable medium distributed separately from the device body in order to provide the program to the user, but also of a recording medium etc. provided to the user in a state in which it is pre-installed in the device body.

[0062] Although the embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment. Furthermore, the effects described in the present embodiment are merely a list of the most preferable effects resulting from the present invention, and the effects of the present invention are not limited to those described in the present embodiment. [Explanation of symbols]

[0063] 1 Information processing system, 10 User terminal, 20 Medication history management server, 30 Voice recognition server, 40 Generative AI service providing server, 50 Network, 151 User interface display control unit (UI display control unit), 152 API set acquisition unit, 153 Voice recognition processing unit, 154 Prompt generation unit, 155 Medication history management system control unit, 171 API set storage unit, 172 Sample information storage unit, 173 Prompt storage unit, 174 Answer information storage unit, 175 Work result storage unit, 800 Information processing device, 811 CPU, 812 ROM, 813 RAM, 814 Bus, 815 Input unit, 816 Output unit, 817 Storage unit, 818 Communication unit, 819 Drive, 820 Imaging unit, 831 Removable media

Claims

1. An information processing system including a terminal device used by a user and a server that executes an operation in response to a request from the terminal device, The terminal device a request receiving means for receiving a task request from a user to the server in natural language; an answer request means for transmitting the contents of the work request received by the request receiving means, interface information of information processing available on the server, and a sample of control information for the server using the interface information as a prompt to a generative AI; A task request sending means for acquiring answer information sent from the generative AI in response to sending the prompt, and sending the answer information as control information to the server; Equipped with The server, an information processing system which executes work in accordance with control information transmitted from said work request transmission means, and transmits work results to said terminal device;

2. The information processing system according to claim 1, characterized in that the terminal device is provided with a work result output means for outputting, in response to the transmission of the control information, the work result transmitted from the server, the control information, and a work request to the server in natural language input by a user when the control information was generated.

3. The answer request means transmits the prompt to the generative AI, including information for specifying an output format of the answer information; The information processing system according to claim 1 or 2, characterized in that the work request sending means obtains the answer information output in the output format from the generative AI and sends the answer information to the server as the control information.

4. The answer request means transmits the prompt to the generative AI, the prompt including information specifying a rule for generating the answer information; The information processing system according to claim 1 or 2, characterized in that the work request sending means obtains the answer information generated in accordance with the rules from the generative AI and sends the answer information to the server as the control information.

5. As a task expressed in natural language, a task name is defined that represents a series of tasks that combine multiple tasks. the request receiving means receives the task name as a task request from a user to the server in natural language; The answer request means transmits the prompt to the generative AI to obtain the answer information for each of a series of tasks defined as the task name, The information processing system according to claim 1 or 2, characterized in that the work request sending means acquires the answer information sent from the generative AI and sends it to the server as the control information for acquiring the work results of each of a series of works defined as the work name.

6. The server, Equipped with the function to link with other systems, The information processing system according to claim 1 or 2, characterized in that when a work request from the user to the server in natural language has content related to the other system, the work is performed in accordance with the control information, and a work related to the other system is also performed, and the work results including the work related to the other system are transmitted to the terminal device.

7. A terminal device constituting an information processing system including a terminal device used by a user and a server that executes an operation in response to a request from the terminal device, a request receiving means for receiving a task request from a user to the server in natural language; an answer request means for transmitting the contents of the work request received by the request receiving means, interface information of information processing available on the server, and a sample of control information for the server using the interface information as a prompt to a generative AI; A task request sending means for acquiring answer information sent from the generative AI in response to sending the prompt, and sending the answer information as control information to the server; A terminal device comprising:

8. An information processing method executed by an information processing system including a terminal device used by a user and a server that executes an operation in response to a request from the terminal device, comprising: The terminal device, a request receiving step of receiving a task request from a user to the server in natural language; a response request step of transmitting the contents of the work request received in the request receiving step, interface information of information processing available on the server, and a sample of control information for the server using the interface information as a prompt to the generative AI; A task request transmission step of acquiring answer information transmitted from the generative AI in response to the transmission of the prompt and transmitting the answer information as control information to the server; Including, The server, An information processing method comprising the steps of: executing a task in accordance with the control information transmitted in said task request transmitting step; and transmitting a task result to said terminal device.

9. A computer constituting a terminal device in an information processing system including a terminal device used by a user and a server that executes operations in response to requests from the terminal device, a request receiving function for receiving a task request from a user to the server in natural language; an answer request function that transmits the contents of the work request received by the request receiving function, interface information of information processing available on the server, and a sample of control information for the server using the interface information as a prompt to the generative AI; A work request transmission function that acquires answer information transmitted from the generative AI in response to the transmission of the prompt and transmits it as control information to the server; A program for realizing the above.

Images