Service document main point presentation device

Abstract

A service document main point presentation device 1 is for presenting a main point DG of an electronic document ED required to be confirmed by a user U who receives service provision, the service document main point presentation device comprising: a service-type recognition unit 12 that recognizes the type of service targeted by the electronic document ED; a model selection unit 13 that selects, according to the recognized type of the service, at least one main point identification model M for use in identifying the main point DG of the electronic document ED from among a larger number of main point identification models MG; and a main point identification unit 14 that inputs the electronic document ED to the selected at least one main point identification model M and identifies the main point DG to be presented to the user U.

Description

Service Document Key Point Presentation Device 【0001】 The present disclosure relates to a service document key point presentation device and the like. 【0002】 Patent Document 1 discloses a contract document checking device that supports the checking of contract documents by a person in charge. For each contract content in the contract document, conditions to be satisfied and conditions that should not be satisfied are preset as check condition information, and based on this, the contract document is automatically checked. When there is contract content for which the conditions to be satisfied are not satisfied, or when there is contract content for which the conditions that should not be satisfied are satisfied, a screen display including the basis is presented to the person in charge. 【0003】 Japanese Patent Application Laid-Open No. 2020-123197 【0004】 In Patent Document 1, for each contract content in the contract document, it is necessary to preset conditions to be satisfied and conditions that should not be satisfied as check condition information, which is cumbersome. 【0005】 The present disclosure has been made in view of such a situation, and an object thereof is to provide a service document key point presentation device and the like that appropriately present the key points of an electronic document for which confirmation by a user receiving a service is required. 【0006】 In order to solve the above problems, a service document key point presentation device according to an aspect of the present disclosure is a service document key point presentation device that presents key points of an electronic document for which confirmation by a user receiving a service is required, and includes a service type recognition unit that recognizes the type of service targeted by the electronic document, and a model selection unit that selects at least one key point identification model to be used for identifying the key points of the electronic document from a larger number of key point identification models according to the recognized service type, and a key point identification unit that inputs the electronic document into the selected at least one key point identification model and identifies the key points to be presented to the user. 【0007】 According to this aspect, the key points of the electronic document can be efficiently identified and presented to the user by the key point identification model selected according to the type of service targeted by the electronic document. 【0008】Another aspect of this disclosure is a service document summary presentation device. This device presents the key points of an electronic document that a user receiving a service is required to confirm, and includes a key point identification unit that takes the electronic document as input to a key point identification model and identifies the key points to be presented to the user. The key point identification model is a generation AI model, and the key point identification unit includes a prompt generation unit that generates prompts to be input along with the electronic document to the key point identification model. Direct access to the prompt generation unit by the service provider is restricted. 【0009】 According to this embodiment, by restricting the service provider's direct access to the prompt generation unit, which directly influences the identification of key points in electronic documents by the key point identification model, it is possible to prevent the service provider's arbitrariness or bias from being excessively reflected in the prompts and / or key point identification results. As a result, the fairness of the key point identification results by the key point identification model can be enhanced. 【0010】 Another aspect of this disclosure is a service document summary presentation device. This device presents the key points of an electronic document that a user receiving a service is required to confirm, and comprises a key point identification unit that inputs the electronic document into a key point identification model and identifies the key points to be presented to the user, and a statistical analysis unit that statistically analyzes a group of electronic documents relating to multiple different services. The key point identification unit refers to the results of the analysis in identifying the key points of the electronic document and identifies items that show a statistically significant difference from the group of electronic documents as key points. 【0011】 According to this embodiment, based on the results of a statistical analysis of a group of electronic documents relating to multiple different services, items showing statistically significant differences can be efficiently identified as the key points of the electronic documents. 【0012】 Furthermore, any combination of the above components, as well as any representations thereof converted into methods, apparatus, systems, recording media, computer programs, etc., are also included in this disclosure. 【0013】This is a schematic functional block diagram of a service document summary presentation device. An example of data stored in the database is schematically shown. A first embodiment of the summary identification model group is schematically shown. A second embodiment of the summary identification model group is schematically shown. An example of a screen in which a user can start checking the summary of an electronic document is shown. An example of a screen for checking the summary of an electronic document is shown. An example of a screen for checking the details of the summary of an electronic document is shown. This is a flowchart schematically showing the flow of automatic generation and confirmation of the summary of an electronic document by the service provider. An example of a processing screen that the service provider can operate and confirm according to the flowchart in Figure 8 is schematically shown. 【0014】 The following describes in detail the forms for implementing this disclosure (hereinafter also referred to as embodiments) with reference to the drawings. In the description and / or drawings, identical or equivalent components, members, processes, etc., are denoted by the same reference numerals, and redundant descriptions are omitted. The scale and shape of the illustrated parts are set for convenience to simplify the description and are not to be interpreted restrictively unless otherwise specified. The embodiments are illustrative and do not limit the scope of this disclosure in any way. Not all features or combinations thereof presented in the embodiments are necessarily essential to this disclosure. For convenience, embodiments are presented by breaking them down into components for each function and / or group of functions that realize them. However, one component in an embodiment may actually be realized by a combination of multiple components as separate entities, and multiple components in an embodiment may actually be realized by a single component as a whole. Furthermore, multiple embodiments and modifications may be disclosed in parallel, and any components of each embodiment and / or modification may be combined in any manner as long as they do not interfere with each other's functions. 【0015】Figure 1 is a schematic functional block diagram of the service document key points presentation device 1 according to this embodiment. The service document key points presentation device 1 comprises an acquisition unit 11, a service type recognition unit 12, a model selection unit 13, a key points identification unit 14, a key points presentation unit 15, a statistical analysis unit 16, a key points re-identification request reception unit 17, and a key points re-identification request control unit 18. Some of these functional blocks can be omitted as long as the service document key points presentation device 1 can achieve at least some of the operations and / or effects described below. 【0016】 These functional blocks are realized through the collaboration of hardware resources such as the computer's central processing unit, memory, input devices, output devices, and peripheral devices connected to the computer, and software that runs using them. Regardless of the type or location of the computer, each of the above functional blocks may be realized using the hardware resources of a single computer, or it may be realized by combining hardware resources distributed across multiple computers. In particular, in this embodiment, it is preferable that some or all of the functional blocks of the service document summary presentation device 1 are realized centrally or distributedly by servers accessible to the service provider SV and user U via a communication network NW such as the Internet, as described later. However, some of the functional blocks of the service document summary presentation device 1 may be realized by computers that are locally accessible to the service provider SV and / or user U. 【0017】 Furthermore, some or all of the functional blocks of the service document summary presentation device 1 may be composed of artificial intelligence (AI). In this case, each functional block may be composed individually by substantially independent AIs, or multiple or all of the functional blocks may be composed by substantially one integrated AI. 【0018】The service document summary presentation device 1 according to this embodiment is a device that presents the summary DG of an electronic document ED that is required to be confirmed by a user U who receives a service from a service provider SV. In order for user U to receive various services from various service providers SV, user U is required to confirm and consent to various electronic documents ED. However, many of these electronic documents ED typically have the characteristics of legal documents, and therefore their contents are complex, difficult to understand, and contain a large amount of information. Generally, user U may, in their priority of receiving the service, consent to an electronic document ED that could legally bind them without carefully examining it. 【0019】 However, this situation is undesirable not only for the user U but also for the service provider SV. The user U may develop distrust towards the service provider SV and the service itself, given that they are forced to agree to a large amount of electronic documents ED that they cannot realistically scrutinize in order to receive the service. Furthermore, even if the service provider SV presents transparent electronic documents ED to the user U, the user U may still feel that the electronic documents ED themselves, which they were forced to agree to without scrutiny, and the consent process were not sufficiently transparent or fair. 【0020】 The service document summary presentation device 1 according to this embodiment solves the above-mentioned problems by automatically identifying the key points DG of an electronic document ED and presenting them to the user U. The user U can efficiently grasp the key points DG without having to examine the electronic document ED itself, and can appropriately decide whether or not to consent to the electronic document ED in order to receive the service. Because such a consent process involves the active participation of the user U in grasping and deciding on the key points DG, the user U is more likely to feel that it is transparent and fair. As a result, it is possible to increase the user U's trust in the service provider SV and the service itself. 【0021】Examples of electronic documents ED that require confirmation or consent from the user U receiving the service include, but are not limited to, terms of service that define the rules and conditions for using the service, privacy policies, personal information protection policies, external transmission policies and other documents concerning the handling of personal information, notations or displays based on laws and regulations such as the Act on Specified Commercial Transactions, return policies, refund policies, and service guides. These electronic documents ED are electronically accessible to the user U via a communication network NW, etc., and legally bind the user U and / or the service provider SV when the user U explicitly consents to or implicitly approves of them when using the service. 【0022】 The acquisition unit 11 acquires one or more electronic documents ED that user U needs to confirm or consent to in order to receive services from service provider SV, and / or various service information SI about the services provided by service provider SV. The service information SI may be interpreted as metadata representing details and supplementary information about the services provided by service provider SV. 【0023】 The electronic documents ED and / or service information SI acquired by the acquisition unit 11 are preferably stored in a centralized database DB. This database DB may aggregate data (including, but not limited to, electronic documents ED and service information SI) of substantially all (or more) services and / or substantially all (or more) service providers SV that utilize the service document summary presentation device 1 according to this embodiment. 【0024】 The acquisition unit 11 may directly acquire electronic documents ED and / or service information SI from the service provider SV, or it may indirectly acquire electronic documents ED and / or service information SI that the service provider SV has uploaded to a website or webpage on the communication network NW. In the latter case, it is preferable that the acquisition unit 11 has prior knowledge of the location on the communication network NW, such as a URL (Uniform Resource Locator) or address, where the electronic documents ED and / or service information SI may be uploaded or updated. 【0025】 Information on the location on the communication network NW where the electronic document ED and / or service information SI can be uploaded or updated may be provided by the service provider SV to the operator of the service document summary presentation device 1 when the service provider SV starts using the service document summary presentation device 1 according to this embodiment. Alternatively, the acquisition unit 11 may be configured by a crawler or the like that visits service-related websites to acquire the latest electronic document ED and / or service information SI. 【0026】 Thus, the acquisition unit 11, which acquires electronic documents ED and / or service information SI from the communication network NW, preferably accesses service-related websites or predetermined URLs at a fixed or variable frequency that does not exceed a predetermined interval (for example, one day or 12 hours) in principle, and periodically checks for updates or other changes in the uploaded electronic documents ED and / or service information SI. 【0027】 If an updated electronic document ED and / or service information SI is found to have no substantial impact on the processing of the service document summary presentation device 1 described later, it is stored in the database DB as the latest version of the electronic document ED and / or service information SI, along with its update date and time. In addition, the summary DG (described later) presented to the user U may be automatically updated according to the updated electronic document ED and / or service information SI. If the database DB has sufficient capacity, it is preferable to also store the electronic document ED and / or service information SI before the update for use by the statistical analysis unit 16, etc., described later. Furthermore, in order for the operator of the service document summary presentation device 1 to easily check the update history of the electronic document ED and / or service information SI, it is preferable to store the electronic document ED and / or service information SI before and after each update in the database DB, associated with each other. 【0028】On the other hand, if there is an update to the uploaded electronic document ED and / or service information SI, and it is determined that this update may have a substantial impact on the processing of the service document summary presentation device 1 described later, the latest version of the electronic document ED and / or service information SI may be stored in the database DB along with the update date and time, and the operator of the service document summary presentation device 1 may also be notified. Upon receiving such a notification or alert, the operator of the service document summary presentation device 1 can examine the contents of the update and quickly take necessary actions such as adjusting or retraining the prompt generation unit 141 and the summary identification model M described later. 【0029】 Furthermore, if the periodic patrol checks by the acquisition unit 11 result in the inability to access the uploaded electronic document ED and / or service information SI, it is preferable that the operator of the service document summary presentation device 1 and the service provider SV are immediately notified. Upon receiving such a notification or alert, the operator of the service document summary presentation device 1 can examine the actual situation and quickly take necessary actions, such as notifying the service provider SV to prompt them to take measures for recovery. 【0030】 The service type recognition unit 12 recognizes the type of service targeted by the electronic document ED, which is acquired by the acquisition unit 11 and stored in the database DB as necessary. The service type recognition unit 12, which may be composed of AI, may recognize the service type by analyzing the electronic document ED itself (typically containing information that suggests the content and type of the service), but it is preferable to recognize the service type based on the service information SI acquired by the acquisition unit 11 and stored in the database DB as necessary. 【0031】Figure 2 schematically shows an example of data stored in a database DB that the service type recognition unit 12 can refer to in order to recognize the service type ST. As mentioned above, the database DB may centrally store service information SI and / or electronic documents ED for each service provided by each service provider SV in a common format. In the example in Figure 2, the service information SI includes ID, service provider name, service name, service segment, special notes, and service URL, and the electronic documents ED include terms of service URL, privacy policy URL, information based on the Act on Specified Commercial Transactions URL, and other electronic document URLs. 【0032】 An ID is a number or symbol arbitrarily assigned in a database to uniquely identify a service. The service provider name is the name of the service provider (SV) that provides the service. The service name is the name of the service provided by the service provider (SV). The service URL represents the location on the communication network (NW) where the service is provided, or where information about the service is provided. 【0033】 A service segment represents the segment, category, type, etc., of services provided by a service provider (SV). Examples of service segments include, but are not limited to, e-commerce (EC), auctions, social networking services (SNS), subscriptions, matching, payments, communication, storage, human resources, bulletin boards, blogs, search, C2C (Consumer to Consumer), business efficiency, crowdsourcing, media, and insurance. 【0034】Special notes refer to special notes concerning the services provided by the service provider SV. In this embodiment, special notes mean any matters other than the service segment that directly or indirectly affect the content of the electronic document ED, such as the terms of use. For example, even if the service segment is the same "EC," the content of the electronic document ED, such as the terms of use, may differ for shopping mall-type EC services (special note "mall"), EC services that substantially sell only single items or single types of products (special note "single item sales"), and EC services that are linked to a point service or point program (special note "point linkage available"). 【0035】 Specifically, in e-commerce services that essentially sell only single items or single types of products, there is a high probability that clauses regarding the recurring purchase of those products will be present in the terms of service as an electronic document (ED). Furthermore, in e-commerce services that are linked to point services or point programs, there is a high probability that clauses regarding point redemption rates and conditions for point accrual will be present in the terms of service as an electronic document (ED). Note that special provisions are not limited to these, and may include any other matters that directly or indirectly affect the content or structure of the electronic document (ED), such as the terms of service. 【0036】 The service segments and / or special notes described above may constitute a service type ST recognized by the service type recognition unit 12. That is, in the example of Figure 2, the service type recognition unit 12 may recognize a set of service segments and special notes as a service type ST. As mentioned above, the service segments and special notes that can constitute a service type ST directly or indirectly affect the content of the electronic document ED, such as the terms of use. As will be described later, in this embodiment, a key point identification model M is constructed for each service type ST that can be recognized by the service type recognition unit 12, which identifies the key points DG of the electronic document ED. Therefore, specific clauses and content that appear in the electronic document ED according to the service type ST can be presented to the user U as key points DG without any omissions. 【0037】The Terms of Service URL is the URL of the Terms of Service for the service. The Privacy Policy URL is the URL of the Privacy Policy for the service. The Specified Commercial Transactions Law Statement URL is the URL of the Specified Commercial Transactions Law Statement for the service. Other Electronic Document URLs are URLs of other electronic documents ED that require confirmation or consent from the user U regarding the service. Thus, a single service may have multiple electronic documents ED, and these sets of electronic documents ED are associated with the service information SI of that single service and stored in the database DB. In addition to or instead of these electronic document ED URLs, the electronic documents ED themselves (i.e., electronic files) may also be stored in the database DB. 【0038】 In the example shown in Figure 2, the URL for the terms of service for service ID "12" is "not found," meaning user U cannot access the terms of service. As mentioned above, such undesirable situations can be quickly detected by the acquisition unit 11 periodically accessing the URL, and immediately notified to the operator of the service document summary presentation device 1, the service provider SV, etc. 【0039】 As will be described later, at least a portion of the service information SI and / or electronic document ED described above may be directly input to the service document summary presentation device 1 when the service provider SV starts using the service document summary presentation device 1 according to this embodiment. 【0040】 In Figure 1, the model selection unit 13 selects at least one key point identification model M from a larger group of key point identification models MG to be used to identify the key points DG of the electronic document ED related to the service, according to the type of service recognized by the service type recognition unit 12. 【0041】 Figure 3 schematically shows a first embodiment of a key point identification model group MG composed of multiple key point identification models M. In this example, one key point identification model M is pre-constructed for each service type ST and each type of electronic document ED recognized by the service type recognition unit 12. 【0042】 Here, the service type ST is determined by the combination of the service segment and special notes exemplified in FIG. 2. In FIG. 3, service type A "EC / Mall" which is the combination of service segment "EC" and special note "Mall", service type B "EC / Single Item Sale" which is the combination of service segment "EC" and special note "Single Item Sale", service type C "EC / With Point Linkage" which is the combination of service segment "EC" and special note "With Point Linkage", and service type D "Auction" which is equal to the service segment "Auction" are exemplified. Also, the types of electronic document ED are exemplified by the terms of use, privacy policy, notations based on the Specified Commercial Transactions Act, and other electronic documents exemplified as URLs in FIG. 2. 【0043】 As schematically exemplified in FIG. 3, the key point identification model M may be constructed individually or one-to-one for all combinations or pairs of the service type ST and the type of electronic document ED. However, for a specific type of electronic document ED, a plurality of different service types ST may share the same key point identification model M. For example, notations based on the privacy policy and the Specified Commercial Transactions Act are considered to have less variation in content due to differences in the service type ST compared to the terms of use, so the key point identification model M may be shared across a plurality of different service types ST. 【0044】 FIG. 4 schematically shows a second embodiment of the key point identification model group MG composed of a plurality of key point identification models M. In the example of this figure, a separate model is pre-constructed for each of the service segment and special notes that define the service type ST in the example of FIG. 2. Note that the example of this figure shows the composition of the key point identification model M for a specific type of electronic document ED (for example, the terms of use). 【0045】For example, for service segments, a main model that constitutes the main part of the key-identification model M is pre-built. In the example in Figure 4, there is one "main model for e-commerce" for the service segment "e-commerce" and one "main model for auctions" for the service segment "auctions". Furthermore, for special notes, submodels that constitute the dependent parts of the key-identification model M are pre-built. In the example in Figure 4, there is one "submodel for malls" for the special note "mall", one "submodel for single-item sales" for the special note "single-item sales", and one "submodel for point linkage" for the special note "point linkage available". 【0046】 The key point identification model M selected by the model selection unit 13 according to the service type ST may be determined by a combination of the main model and sub-model as described above (or by the main model only). 【0047】 Specifically, when the aforementioned service type A "EC / Mall" is recognized by the service type recognition unit 12, the model selection unit 13 selects the set of "Main model for EC" and "Sub-model for Mall" as the key-identification model M. When the aforementioned service type B "EC / Single item sales" is recognized by the service type recognition unit 12, the model selection unit 13 selects the set of "Main model for EC" and "Sub-model for single item sales" as the key-identification model M. When the aforementioned service type C "EC / Point linkage enabled" is recognized by the service type recognition unit 12, the model selection unit 13 selects the set of "Main model for EC" and "Sub-model for point linkage enabled" as the key-identification model M. When the aforementioned service type D "Auction" is recognized by the service type recognition unit 12, the model selection unit 13 selects the "Main model for auction" as the key-identification model M. 【0048】According to the above embodiments, for service types A to C where the service segment is common as "EC", one main model, the "main model for EC" can be shared (schematically shown by the double-headed arrows in FIG. 4), thus reducing the number and total size of the models to be constructed in advance. Note that the service segment and specific matters that define the service type ST may each be further subdivided, hierarchized, paralleled, and / or complicated, and accordingly, model modules (small models modularized in a combinable manner as exemplified in FIG. 4) that are components of the key point identification model M may be constructed in advance. 【0049】 In FIG. 1, the key point identification unit 14 inputs the electronic document ED acquired by the acquisition unit 11 into at least one key point identification model M selected by the model selection unit 13, and identifies the key points DG presented to the user U. Here, the key point identification model M selected by the model selection unit 13 is, for example, a trained model that has learned comprehensive learning data and teacher data so as to appropriately identify the key points DG of a specific type of electronic document ED (e.g., terms of use) related to a service of a specific service type ST (recognized by the service type recognition unit 12) by an AI having a learning function such as machine learning (ML: Machine Learning). 【0050】The key point identification model M, which is an AI model or AI / ML model as described above, may be configured as a generating AI model. In this case, it is preferable that the key point identification unit 14 includes a prompt generation unit 141 that generates a prompt P to be input together with the electronic document ED to at least one key point identification model M selected by the model selection unit 13. The prompt generation unit 141 recognizes the service type ST recognized by the service type recognition unit 12 and the key point identification model M selected accordingly by the model selection unit 13, and then automatically generates a prompt P according to a predetermined logic or algorithm that causes the key point identification model M, which has been trained on the specified service type ST, to accurately extract key points DG from the electronic document ED acquired by the acquisition unit 11. As will be described later, since the service provider SV's access to the prompt generation unit 141 is restricted or prohibited, the fairness of the key point DG identification results by the prompt P and the key point identification model M can be enhanced. 【0051】 The model selection unit 13 and the key point identification unit 14 described above may handle each of the multiple electronic documents ED (terms of service, privacy policy, legal disclaimers such as the Act on Specified Commercial Transactions, return policy, refund policy, service guide, etc.) relating to a particular service individually, or they may handle substantially all of the electronic documents ED relating to a particular service as a group in an integrated manner. In the latter case, the model selection unit 13 selects all key point identification models M or all model modules corresponding to all electronic documents ED (see Figure 2) related to the service to be identified as key point. The key point identification unit 14 then takes all the electronic documents ED relating to the service and a prompt P as input to the integrated key point identification model M, which is a combination of these multiple key point identification models M or multiple model modules, and outputs the key points DG of the electronic documents group. 【0052】The key points presentation unit 15 presents the key points DG of the electronic document ED, identified by the key points identification unit 14 and the key points identification model M, to the user U who is seeking to receive a service from the service provider SV. For example, the key points presentation unit 15 displays the key points DG of the electronic document ED, identified by the key points identification unit 14 and the key points identification model M, on the screen of an information processing device (not shown) such as a smartphone or personal computer used by the user U. 【0053】 Figure 5 shows an example of a screen where user U can begin reviewing the key points DG of an electronic document ED, such as the terms of service. This screen allows user U to enter personal information such as their name, address, and contact information, and then agree to the terms of service for the electronic document ED (by pressing the button at the bottom of the screen that displays the text "Agree and confirm entered information"). In this figure, the terms of service for the services of "YYY Corporation" are used as an example of the electronic document ED to be reviewed. The full text of the terms of service that user U must review or agree to (only a portion of the beginning is displayed in Figure 5) is shown in a scrollable sub-screen at the bottom of the screen. In addition, a button 151 to transition to the electronic document ED key points DG review screen is highlighted in a prominent location around the terms of service sub-screen. In the example in Figure 5, the text "Review key points of the terms" is displayed on the transition button 151 to smoothly guide user U to the electronic document ED key points DG review screen. 【0054】 When user U presses or selects the transition button 151, the user transitions to a screen to confirm the key points DG of the electronic document ED, as illustrated in Figure 6. This summary screen lists a few key points (five in the example in Figure 6) that should draw user U's attention in the electronic document ED, and each summary is displayed in a list of concise text and diagrams. By simply viewing this summary screen, user U can efficiently and accurately grasp the key points DG of the electronic document ED. 【0055】The number and content of key points displayed on this summary screen may be uniform for all electronic documents ED of the same type, or they may vary adaptively depending on the service information SI including service type ST, the type of electronic document ED, and the key point identification model M which can be customized for them. Furthermore, these key points (items such as "About Money") may be prepared in advance as input data for the key point identification model M, or they may be automatically generated as output of the key point identification model M. 【0056】 When user U presses or selects the details confirmation button 152 (which displays the text "View Details"), the user is redirected to a details confirmation screen as illustrated in Figure 7. This figure shows an example of the details confirmation screen for "2. Regarding Rights," one of the five main points in Figure 6. The content, which was displayed in concise text in Figure 6, is displayed in more detail in Figure 7. However, these details are also summaries of the articles of the electronic document ED and are still interpreted as the main points DG of the electronic document ED. 【0057】 The parentheses at the end of each sentence indicate the corresponding clause number in the electronic document ED (Terms of Use). For example, "(9)" indicates that the basis for this summary is found in Article 9 of the electronic document ED. When User U clicks or selects the original text confirmation button 153 (which displays the text "View Original Text"), they can transition to a screen (not shown) where they can view the original text of the corresponding clauses in the electronic document ED (in the example in Figure 7, Articles 9, 13, and 27). Furthermore, when User U clicks or selects the previous item confirmation button 154 (which displays the text "Previous"), they can transition to a detailed confirmation screen (not shown) for the previous main point, "1. About Money". In addition, when User U clicks or selects the next item confirmation button 155 (which displays the text "Next"), they can transition to a detailed confirmation screen (not shown) for the next main point, "3. About Privacy". 【0058】According to this embodiment, the key points DG of the electronic document ED can be efficiently identified and presented to the user U by a key point identification model M selected according to the type of service to which the electronic document ED is intended. 【0059】 In Figure 1, a statistical analysis unit 16 may be provided to statistically analyze a group of electronic documents ED stored in the database DB. Preferably, the statistical analysis unit 16 statistically analyzes a group of electronic documents ED relating to multiple different services of the same type. The type here may be the service type ST exemplified in Figure 2, or it may be the service segments and / or special notes that are components of the service type ST. Preferably, such a statistical analysis unit 16 can statistically grasp the overall trend of electronic documents ED of various service providers SV relating to a particular type of service (e.g., EC). 【0060】 In identifying the key points DG of one electronic document ED relating to the specified type of service, the key point identification unit 14 may refer to the results of statistical analysis by the statistical analysis unit 16 and identify as key points items that show a statistically significant difference compared to a group of electronic documents relating to services from other service providers SV of the specified type. 【0061】 For example, the statistical analysis unit 16 statistically analyzes a group of refund terms as electronic documents ED related to services of types such as "EC" and "EC / Mall". The target of this statistical analysis is all services registered as "EC" and / or "Mall" in the example in Figure 2 (the service provider SV is not a factor, but it is preferable to exclude the services themselves that the electronic documents ED targeted for key point identification target cover, and the service providers SV of those services). In this case, the refund terms as electronic documents ED may be registered in "Other Electronic Document URLs" or may be posted as part of the terms of service etc. registered in "Terms of Service URLs". 【0062】The statistical analysis unit 16 may statistically analyze the group of refund policies stored in such a database DB and create statistical data, for example, on the period during which refunds are possible after purchase (refundable period). Based on this statistical data, the statistical analysis unit 16 may calculate statistical values ​​such as the mean, median, mode, standard deviation, and variance of the refundable period (e.g., in days). 【0063】 The key point identification unit 14 compares these statistical values ​​with the refund period stipulated in the refund terms for the electronic document ED subject to key point identification, and if a statistically significant difference is found, it identifies the refund period as a key point DG of the refund terms. A statistically significant difference may be determined, for example, based on the mean and standard deviation of the statistical data. Specifically, if the refund period stipulated in the refund terms for the electronic document ED subject to key point identification is n standard deviations (where n is any positive number, for example, 1 or more, 2 or more, or 3 or more) or more away from the mean of the refund period in the statistical data, the refund period may be considered statistically significant and identified as one of the key points DG by the key point identification unit 14. 【0064】 The key points presentation unit 15 may present the key points DG of the electronic document ED, identified by the key points identification unit 14 and the key points identification model M, to the service provider SV for confirmation before presenting them to the user U. For example, the key points presentation unit 15 may display the key points DG of the electronic document ED, identified by the key points identification unit 14 and the key points identification model M, on the screen of an information processing device (not shown) such as a smartphone or personal computer used by the service provider SV. 【0065】 The Key Points Re-Identification Request Reception Unit 17 may receive a request from a service provider SV, who has received a proposal for the key points DG of the electronic document ED relating to its own service, for the key points identification unit 14 to re-identify the key points DG of the electronic document ED. The Key Points Re-Identification Request Control Unit 18 controls the key points re-identification request from the service provider SV. The Key Points Re-Identification Request Control Unit 18 may include, for example, a key points re-identification request restriction unit 181 and a key points re-identification request restriction release unit 182. 【0066】The Key Point Re-identification Request Restriction Unit 181 imposes arbitrary restrictions, such as limits on the number of requests or content restrictions, on the requests for re-identification of the key points DG of the electronic document ED that the Key Point Re-identification Request Reception Unit 17 can receive from the service provider SV. If the service provider SV could make unlimited requests for re-identification of key points, the arbitrariness or bias of the service provider SV would be excessively reflected in the key point identification results by the key point identification unit 14, and there is a risk that the key points DG could not be presented fairly to the user U. Therefore, in this embodiment, the Key Point Re-identification Request Restriction Unit 181 imposes restrictions on the key point re-identification requests made by the service provider SV, thereby increasing the fairness of the key point identification results by the key point identification unit 14 and / or the key point identification model M. 【0067】 If the number of key point re-identification requests made by the service provider SV reaches the upper limit set by the key point re-identification request restriction unit 181, the service provider SV may be prohibited from making key point re-identification requests for the electronic document ED, at least temporarily. The key point re-identification request restriction release unit 182 may release or disable the restriction activated by the key point re-identification request restriction unit 181 when predetermined release conditions are met. Examples of predetermined release conditions include the passage of a predetermined time after the restriction was activated, or the operator of the service document key point presentation device 1 or the key point re-identification request control unit 18 itself taking action to release the restriction. In any case, direct access by the service provider SV to the key point re-identification request restriction unit 181 and the key point re-identification request restriction release unit 182 (i.e., the key point re-identification request control unit 18) is restricted, thereby ensuring fairness in the key point identification results by the key point identification unit 14 and / or the key point identification model M. 【0068】The key point identification unit 14, in response to a key point re-identification request from the service provider SV that was received by the key point re-identification request reception unit 17 and not restricted by the key point re-identification request restriction unit 181, causes at least one key point identification model M selected by the model selection unit 13 to output a key point DG different from the already output key point DG. This different key point DG is then presented to the service provider SV by the key point presentation unit 15. In this way, the service provider SV can compare multiple different patterns of key point DG output from substantially the same key point identification model M and ultimately select the key point DG that should be presented to the user U. 【0069】 Specifically, the prompt generation unit 141 in the key point identification unit 14 may generate a prompt P different from the prompt P already input to at least one key point identification model M selected by the model selection unit 13, in response to a key point re-identification request from a service provider SV that has been received by the key point re-identification request reception unit 17 and not restricted by the key point re-identification request restriction unit 181. This different prompt P is input to the key point identification model M together with the electronic document ED, and a key point DG different from the key point DG already output from the key point identification model M is output. In this way, multiple different patterns of prompt P that are input to the key point identification model M, which is substantially the same generating AI model, may be prepared in advance by the operator of the service document key point presentation device 1, or they may be autonomously generated by the prompt generation unit 141, which may be an AI / ML function unit. 【0070】 In the initial identification of the key points DG of the electronic document ED as described above, and in subsequent re-identifications, direct access by the service provider SV to the key point identification unit 14 and / or the prompt generation unit 141 is restricted, thereby ensuring the fairness of the key point identification results by the key point identification unit 14 and / or the key point identification model M. 【0071】 Figure 8 is a schematic flowchart illustrating the process of automatic generation and verification of the key points DG of the electronic document ED by the service provider SV. In the flowchart, "S" represents a step or process. Figure 9 schematically shows an example of a processing screen that the service provider SV can operate and verify according to the flowchart in Figure 8. 【0072】 In S1, the service provider SV inputs the information necessary for the automatic generation of the main points DG of the electronic document ED to the acquisition unit 11. Examples of the information input in S1 include service information SI and the electronic document ED, as illustrated in Figures 1 and 2. Alternatively, the acquisition unit 11 may automatically identify the remaining information based on some of the information input by the service provider SV and automatically input it into the corresponding input fields. 【0073】 In the example processing screen shown in Figure 9, the service provider SV inputs the following: service URL (service information SI), terms and conditions URL (electronic document ED), service name (service information SI), terms and conditions name (electronic document ED), service type (service type ST, which is part of the service information SI), service provider SV's email address (service information SI), desired URL for the key points DG (location for displaying the key points DG to user U), and notes for the consumer (user U). Here, the service type field may be selectable from the default service type ST as exemplified in Figures 2 to 4. The information input by the service provider SV is not limited to these and can be arbitrarily set in the service document key points display device 1. For example, the service provider SV may be required to upload the electronic document ED as an electronic file in place of or in addition to the terms and conditions URL and / or terms and conditions name. 【0074】 In S2, the service type recognition unit 12 recognizes the service type ST entered in the service type field as the type of service targeted by the electronic document ED (in the example in Figure 9, "YYY Service Terms of Use"). In S3, the model selection unit 13 selects at least one key point identification model M from the key point identification model group MG to be used to identify the key points DG of the electronic document ED, according to the service type ST recognized in S2. 【0075】When the service provider SV presses or selects the key point identification button 170 in Figure 9 (which displays the text "Show preview"), the process proceeds to S4, where the key point identification unit 14 inputs the electronic document ED (and prompt P in the example of Figure 1) entered in S1 into the key point identification model M selected in S3, extracts the key points DG to be presented to the user U, and presents them to the service provider SV. In the example of Figure 9, the key point identification unit 14 presents the service provider SV with a preview of the key points DG, including a small number of main points (points 1 to 5) corresponding to Figure 6, in the preview display area on the right side of the screen. 【0076】 When the service provider SV, after reviewing the preview of the key points DG presented in S4, presses or selects the key points re-identification request button 171 (which displays the text "Remake") in Figure 9, S5 determines "Yes" and proceeds to S6 (the key points re-identification request receiving unit 17 receives the key points re-identification request from the service provider SV). 【0077】 In S6, the Key Point Re-specification Request Control Unit 18 determines whether the restriction on key point re-specification requests from the service provider SV by the Key Point Re-specification Request Restriction Unit 181 is enabled. For example, if the number of key point re-specification requests from the service provider SV has reached the upper limit set by the Key Point Re-specification Request Restriction Unit 181, the restriction by the Key Point Re-specification Request Restriction Unit 181 is enabled, and in S6, it is determined to be "Yes" and the process proceeds to S7. 【0078】On the other hand, if the number of key point re-identification requests from the service provider SV has not reached the upper limit set by the key point re-identification request limiting unit 181, the limit by the key point re-identification request limiting unit 181 is invalidated, and in S6, it is determined to be "No" and the process returns to S4. In such subsequent S4s, the key point identification unit 14, in response to the key point re-identification request received in S5, causes the key point identification model M selected in S3 (which is substantially the same as the one used in previous S4s) to display a key point DG different from the key point DG output in previous S4s in the preview display area in Figure 9. As mentioned above, in subsequent S4s, the prompt generation unit 141 may generate a prompt P different from the prompt P used in previous S4s and input it to the key point identification model M. 【0079】 In S7, the Key Points Re-specification Request Restriction Release Unit 182 determines whether the restriction on the Key Points Re-specification Request from the Service Provider SV has been released, and the Key Points Re-specification Request Control Unit 18 determines whether or not the restriction on the Key Points Re-specification Request from the Service Provider SV has been released. If the restriction on the Key Points Re-specification Request from the Service Provider SV has not been released, S7 is determined to be "No" and the process proceeds to S8. On the other hand, if the restriction on the Key Points Re-specification Request from the Service Provider SV has been released, S7 is determined to be "Yes" and the process returns to S4. 【0080】 If the key point re-identification request button 171 is not operated in S5 ("No"), or if "No" is determined in S7, the process proceeds to S8. In S8, the service provider SV selects one of the one or more key point DG patterns created in one or more S4s, and then presses or selects the key point determination button 172 (which displays the text "Issue Tag") in Figure 9. The selected key point DG is then finally determined to be presented to the user U. The final determined key point DG may be provided to the service provider SV as a tag to be embedded in the webpage of the desired installation URL entered by the service provider SV on the left side of Figure 9. 【0081】The final decision result of the service provider SV in S8, as described above, can be used as training data or training data for evaluating and improving the key point identification model M and prompt P. For example, the key point identification model M and / or prompt P that contributed to the generation of the key point DG ultimately selected by the service provider SV will be highly evaluated, while the key point identification model M and / or prompt P that contributed to the generation of the key point DG that was not ultimately selected by the service provider SV will be lowly evaluated, and appropriate adjustments such as retraining may be made as needed. 【0082】 The present disclosure has been described above based on embodiments. Various modifications are possible for each component and each combination of processes in the exemplary embodiments, and it will be obvious to those skilled in the art that such modifications are included in the scope of the present disclosure. 【0083】 In the embodiments described above, a key-identification model M was prepared for each service type ST. However, a single, essentially general-purpose key-identification model M capable of handling any type of service may be provided. In this case, the service type recognition unit 12 and the model selection unit 13 may not be provided. Furthermore, in the embodiments described above, the key-identification model M was configured as a generation AI model. However, it may be configured as any type of AI model, not limited to a generation AI model. 【0084】 The configuration, operation, and function of each device and method described in the embodiments can be realized by hardware resources or software resources, or by the cooperation of hardware resources and software resources. Hardware resources include, for example, processors, ROMs, RAMs, and various integrated circuits. Software resources include, for example, operating systems and application programs. 【0085】 This disclosure relates to a service document summary presentation device, etc. 【0086】1 Service document key points presentation device, 11 Acquisition unit, 12 Service type recognition unit, 13 Model selection unit, 14 Key point identification unit, 15 Key point presentation unit, 16 Statistical analysis unit, 17 Key point re-identification request reception unit, 18 Key point re-identification request control unit, 141 Prompt generation unit, 181 Key point re-identification request restriction unit, 182 Key point re-identification request restriction release unit.

Claims

1. A service document summary presentation device that presents the essential points of an electronic document that are required to be confirmed by a user receiving a service, comprising: a service type recognition unit that recognizes the type of service to which the electronic document pertains; a model selection unit that selects at least one summary identification model to be used for identifying the essential points of the electronic document from a larger number of summary identification models, according to the recognized service type; and a summary identification unit that inputs the electronic document into the selected at least one summary identification model and identifies the essential points to be presented to the user.

2. The service document key point presentation device according to claim 1, wherein the key point identification model is a generation AI model, and the key point identification unit includes a prompt generation unit that generates a prompt to be input together with the electronic document to the selected at least one key point identification model.

3. A service document key point presentation device according to claim 2, comprising a key point re-identification request receiving unit that receives a request from a service provider providing the service to re-identify the key points of the electronic document by the key point identification unit, and the prompt generation unit generates a prompt different from the prompts already entered in the selected at least one key point identification model in response to the re-identification request.

4. The service document summary presentation device according to claim 3, wherein direct access by the service provider to the prompt generation unit is restricted.

5. The service document summary presentation device according to claim 4, wherein the number of requests for re-identification of the key points of the electronic document that the key point re-identification request receiving unit can receive from the service provider is limited.

6. A service document key point presentation device according to any one of claims 1 to 5, comprising a statistical analysis unit for statistically analyzing a group of electronic documents relating to multiple different services of the same type, wherein the key point identification unit refers to the results of the analysis in identifying the key points of the electronic documents and identifies items that show a statistically significant difference from the group of electronic documents as key points.

7. A service document summary presentation device that presents the key points of an electronic document that are required to be confirmed by a user receiving a service, comprising: a key point identification unit that inputs the electronic document into a key point identification model and identifies the key points to be presented to the user, wherein the key point identification model is a generation AI model, and the key point identification unit comprises a prompt generation unit that generates prompts to be input together with the electronic document into the key point identification model, and direct access to the prompt generation unit by the service provider providing the service is restricted.

8. A service document summary presentation device that presents the key points of an electronic document that is required to be confirmed by a user receiving a service, comprising: a key point identification unit that inputs the electronic document into a key point identification model and identifies the key points to be presented to the user; and a statistical analysis unit that statistically analyzes a group of electronic documents relating to multiple different services, wherein the key point identification unit refers to the results of the analysis in identifying the key points of the electronic document and identifies items that show a statistically significant difference from the group of electronic documents as key points.

Images