Information analysis device, information analysis method, and program

Abstract

In order to enable people to correctly interpret spread information, this information analysis device comprises: a collection unit that collects information spread in social media; an analysis unit that analyzes a narrative of a sender of the collected information; and an output unit that outputs analysis data including an analysis result of the narrative of the sender of the information.

Description

Information analysis device, information analysis method, and program

[0001] This disclosure relates to an information analysis device, an information analysis method, and a program.

[0002] With the proliferation of social networking services (SNS), information manipulation known as cognitive warfare has emerged. Those who orchestrate cognitive warfare use social media to spread misinformation with the aim of exaggerating people's perceptions. Intentionally disseminated false information is called disinformation. People whose perceptions have been exaggerated by the spread of disinformation may develop hostile feelings towards specific individuals, organizations, or nations. The spread of such disinformation poses a risk of hindering people's free and accurate decision-making and undermining the very foundations of democracy.

[0003] Patent Document 1 discloses an information processing device for detecting fake news. The device in Patent Document 1 collects information from posted comments, including links to news articles, and clusters the comments based on the words contained in them. Based on the clustering results, the device in Patent Document 1 derives information indicating the likelihood that a news article is fake news. The device in Patent Document 1 derives a higher likelihood of being fake news the more comments are included in the clustered cluster.

[0004] Patent No. 7284196

[0005] Various organizations and groups conduct fact-checking against misinformation. However, fact-checking individual pieces of information alone makes it difficult to infer the attacker's true intentions. As a result, situations arose where the attacker's intended outcomes could occur, such as causing domestic unrest or enhancing one's own country's reputation while damaging the reputation of others. Furthermore, if false public opinion was formed under such circumstances, there was a possibility that citizens would take actions intended by the attacker. Therefore, it is necessary to predict the attacker's true intentions from the individual pieces of information being disseminated and take appropriate countermeasures.

[0006] According to the method described in Patent Document 1, it is possible to determine whether news is fake news or not based on the presented reliability level. However, the method described in Patent Document 1 could not predict the attacker's true objective from the individual pieces of information being disseminated and take appropriate countermeasures.

[0007] The purpose of this disclosure is to provide an information analysis device, an information analysis method, and a program that enable people to correctly interpret disseminated information.

[0008] An information analysis device according to one aspect of this disclosure comprises a collection unit that collects information disseminated on social media, an analysis unit that analyzes the narrative of the information sender, and an output unit that outputs analysis data including the results of the analysis of the information sender's narrative.

[0009] In one aspect of the information analysis method of this disclosure, a computer collects information disseminated on social media, analyzes the narrative of the information sender, and outputs analysis data including the results of the analysis of the information sender's narrative.

[0010] A program in one aspect of this disclosure causes a computer to perform the following processes: collecting information disseminated on social media; analyzing the narrative of the information sender; and outputting analytical data including the results of the analysis of the information sender's narrative.

[0011] This disclosure makes it possible to provide an information analysis device, an information analysis method, and a program that enable people to correctly interpret disseminated information.

[0012] This is a block diagram showing an example of the configuration of the information analysis device in this disclosure. This is a conceptual diagram showing an example of information extraction by the extraction unit of the information analysis device in this disclosure. This is a conceptual diagram showing an example of information classification by the classification unit of the information analysis device in this disclosure. This is a table showing an example of information classified by the classification unit of the information analysis device in this disclosure. This is a table showing an example of evaluation results by the evaluation unit of the information analysis device in this disclosure. This is a table showing an example of parameters referenced by the analysis unit of the information analysis device in this disclosure. This is a conceptual diagram showing an example of information analysis by the analysis unit of the information analysis device in this disclosure. This is a conceptual diagram showing an example of the display of information output from the output unit of the information analysis device in this disclosure. This is a flowchart showing an example of the operation of the information analysis device in this disclosure. This is a conceptual diagram to explain an example of the diffusion pattern of information that is the target of analysis by the information analysis device in this disclosure. This is a conceptual diagram to explain an example of a response in response to the analysis results of the information analysis device in this disclosure. This is a conceptual diagram to explain an example of a response in response to the analysis results of the information analysis device in this disclosure. This is a conceptual diagram to explain an example of a response in response to the analysis results of the information analysis device in this disclosure. This is a conceptual diagram to explain an example of the diffusion pattern of information that is the target of analysis by the information analysis device in this disclosure. This is a conceptual diagram to explain an example of a response in response to the analysis results of the information analysis device in this disclosure. This is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. This is a conceptual diagram illustrating an example of the diffusion pattern of the information being analyzed by the information analysis device in this disclosure. This is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. This is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. This is a block diagram illustrating an example of the configuration of the information analysis device in this disclosure. This is a flowchart illustrating an example of the operation of the information analysis device in this disclosure. This is a block diagram illustrating an example of the hardware configuration for performing the processing in this disclosure.

[0013] The embodiments for carrying out this disclosure will be described below with reference to the drawings. In this disclosure, the drawings used in the description of each embodiment are associated with one or more embodiments. Also, the elements included in each drawing may correspond to one or more embodiments. The embodiments described below have technically preferred limitations for carrying out this disclosure, but the scope of the disclosure is not limited thereto. In all the drawings used in the description of the embodiments below, the same parts are denoted by the same reference numerals unless there is a specific reason not to. In the embodiments below, repeated descriptions of similar configurations and operations may be omitted. The direction of the arrows in the drawings is an example of the flow of signals, data, etc., and does not limit the flow of signals, data, etc.

[0014] (First Embodiment) First, the information analysis device in the first embodiment will be described. The information analysis device in this embodiment collects information disseminated via social media such as SNS (Social Networking Service).

[0015] The information analysis device of this embodiment analyzes the narrative of collected information. A narrative refers to a story or explanation that structures and conveys a series of events or experiences from a specific perspective. Narratives play an important role in the transmission and understanding of information and in forming specific views and beliefs. Information providers combine facts and interpretations to create a consistent and meaningful context. Typical information providers aim to disseminate correct facts and ensure correct interpretation. However, some information providers are attackers who spread malicious information for specific purposes.

[0016] Dissemination of information can include false information. False information includes disinformation and misinformation. Disinformation refers to false or misleading malicious information that is intentionally created and disseminated. Disinformation spreaders intentionally spread false information for a specific purpose. Disinformation is malicious. Disinformation is often created and systematically disseminated for a specific purpose, such as exercising political influence or damaging the reputation of competitors. Misinformation refers to false information that is spread unintentionally. Disseminators of misinformation are unaware that the information is false. Misinformation is not malicious and arises from misunderstanding or ignorance. Misinformation is shared with many people without verifying its accuracy. Dissemination of information can include malicious information. Malinformation is malicious information that, while based on facts, distorts context or encourages misinterpretation, leading the recipient astray. Malinformation is disseminated with the intention of causing misunderstanding or harm. In the following, malicious information such as false information and misinformation will also be referred to as malicious information.

[0017] (Configuration) Next, an example of the configuration of the information analysis device in this embodiment will be described with reference to the drawings. Figure 1 is a block diagram showing an example of the configuration of the information analysis device in this disclosure. The information analysis device 10 comprises a collection unit 11, an extraction unit 12, a classification unit 13, an evaluation unit 15, an analysis unit 16, and an output unit 17. The information analysis device 10 is connected to a communication network (not shown) such as the Internet or an intranet. The information analysis device 10 is also connected to a terminal device (not shown) used by an administrator who performs information analysis.

[0018] The collection unit 11 collects information disseminated on social media via a communication network. The collection unit 11 stores the collected information in a storage unit (not shown). The collected information includes attribute information and metadata indicating the time of transmission, transmission medium, and sender of the information. The collected information may also include attribute information and metadata other than the time of transmission, transmission medium, and sender. The collection unit 11 may be configured to collect information stored in a database. For example, the collection unit 11 collects information disseminated on social media via an API (Application Programming Interface). The social media platforms from which information is collected can be arbitrarily set. The collection unit 11 may be configured to collect information that has been disseminated beyond a certain criterion. For example, the criterion may be set based on numerical values ​​of reach indicators such as the number of reactions, reach, impressions, and hashtag usage frequency.

[0019] Disseminated information can contain false information. False information can include misinformation and misinformation. Furthermore, disseminated information can also contain malicious information. Misinformation and malicious information are considered malicious information. Malicious information can include information that affects international relations. For example, malicious information includes information that is disseminated with a specific purpose regarding a particular target such as a country, region, organization, group, or facility. If malicious or misinformation is disseminated to the public and misinterpreted, readers may take incorrect actions. Therefore, it is necessary to make it known to the public that malicious and misinformation is information that is not factual.

[0020] In this embodiment, the collection unit 11 is configured to collect false information contained in the disseminated information. For example, the collection unit 11 may be configured to collect malicious information in addition to false information. With this configuration, by analyzing malicious information including false information and malicious information, it is possible to respond to information that has been disseminated for a specific purpose. Alternatively, the collection unit 11 may be configured to collect misinformation in addition to false information. With this configuration, by analyzing misinformation in addition to false information, it is possible to respond to misinformation that has been mistakenly disseminated and is not factual.

[0021] The extraction unit 12 extracts false information from the collected information. For example, the extraction unit 12 extracts information containing specific terms or specific events as false information. Specific terms include words and phrases that indicate specific targets or purposes. For example, specific targets include names of countries, regions, organizations, groups, facilities, etc., that are likely to be targeted by malicious information attacks. For example, specific targets include names of countries, regions, organizations, groups, facilities, etc., that tend to be disguised as sources of malicious information. Specific events include terms and sentences related to actions, behaviors, or phenomena that would be problematic if disseminated to the public. For example, specific events include international issues, domestic issues, environmental issues, wars, conflicts, armaments, elections, incidents, etc. Specific terms and specific events can be pre-registered in the memory unit (not shown). The extraction unit 12 may be configured to extract false information by referring to international or domestic norms. In that case, the extraction unit 12 may be configured to refer to an external database. Note that if the information collected by the collection unit 11 is limited to false information, the extraction unit 12 may be omitted.

[0022] The extraction unit 12 may be configured to extract false information from disseminated information in accordance with fact-checking standards. In fact-checking, standards are set to be followed in order to extract false information. There are no particular limitations on the standards set in fact-checking, as long as false information can be extracted from disseminated information. For example, the standards are 12 rules published by the European Fact-Checking Standard Networks (EFCSN). In this case, the extraction unit 12 performs fact-checking in accordance with the 12 rules published by the European Fact-Checking Standard Networks (EFCSN). For example, the standards are five principles that certification bodies of the International Fact-Checking Network (IFCN) adhere to. In this case, the extraction unit 12 performs fact-checking in accordance with the five principles that certification bodies of the International Fact-Checking Network adhere to. For example, the standards may be those published by a Japanese fact-checking organization. In this case, the extraction unit 12 performs fact-checking in accordance with the standards published by the Japanese fact-checking organization.

[0023] For example, the extraction unit 12 extracts false information from disseminated information by fact-checking via an LLM (Large Language Models) system (not shown) that performs processing using a large language model. The LLM system is a system that performs processing using a large language model. A large language model (also called a model) is a deep learning model trained using a large language dataset. The LLM system uses the large language model to output text information corresponding to the content of text information composed of natural language. For example, the LLM system outputs an answer in response to a question input. The LLM system may be a model capable of inputting and outputting images and audio. For example, the LLM system is a system that can be used via an API. The LLM system may be configured to use a dedicated model built to extract false information from disseminated information by fact-checking. There are no limitations on the type of large language model used by the LLM system or the location where the LLM system is located, as long as it can be accessed from the information analysis device 10. For example, the LLM system may be configured inside the information analysis device 10.

[0024] Figure 2 is a conceptual diagram showing an example of information extraction by the extraction unit of the information analysis device in this disclosure. In the example in Figure 2, the extraction unit 12 is connected to the LLM system 100 via API. The extraction unit 12 sends prompt P1 to the LLM system 100 via API. Prompt P1 is input to the LLM system 100. The LLM system 100 performs processing using a large-scale language model and outputs a response A1 corresponding to prompt P1. Prompt P1 includes instructions, norms, and disseminated information. The instructions include the text information, "Extract false information from the information below according to the following norms." Preferably, the instructions are set so that a desired output can be obtained. The norms are documents relating to norms used in fact-checking. The disseminated information is the target of false information extraction. The LLM system 100 extracts false information from the disseminated information according to the norms. Response A1 includes the text information, "Information D and Information J are false information." In this case, the extraction unit 12 extracts information D and information J as false information. The prompt P1 and response A1 shown in Figure 2 are just one example. For example, the prompt for setting the norm in the LLM system 100 and the prompt for extracting false information from the disseminated information may be separated. Also, prompt P1 may include instructions to extract not only false information, but also malicious information or misinformation.

[0025] The classification unit 13 classifies the extracted false information. The classification unit 13 analyzes whether the extracted false information is malicious from the perspective of the 5W1H: "When, Where, Who, What, Why, How". Of the 5W1H perspectives, "When" indicates the timing when the false information was transmitted. Of the 5W1H perspectives, "Where" indicates the social networking service on which the false information was transmitted. Of the 5W1H perspectives, "Who" indicates the name of the entity (account) that transmitted the false information. Of the 5W1H perspectives, "What" indicates the content contained in the false information. Of the 5W1H perspectives, "Why" indicates the purpose for which the false information was transmitted. Of the 5W1H perspectives, "How" indicates the format (text, images, etc.) of the transmitted false information. For example, the classification unit 13 categorizes the false information analyzed from the perspective of the 5W1H according to pre-set criteria. For example, the classification unit 13 may be configured to classify the extracted false information according to the degree of malicious intent.

[0026] For example, the classification unit 13 classifies the extracted false information via an LLM system that performs processing using a large-scale language model. The classification unit 13 may be configured to classify not only false information but also malicious information and harmful information. Alternatively, the classification unit 13 may be configured to classify correct information.

[0027] Figure 3 is a conceptual diagram showing an example of information classification by the classification unit of the information analysis device in this disclosure. In the example in Figure 3, the classification unit 13 is connected to the LLM system 100 via API. The classification unit 13 sends prompt P2 to the LLM system 100 via API. Prompt P2 is input to the LLM system 100. Prompt P2 includes an instruction and the content of false information. The instruction includes the text information, "Classify the following false information from the perspective of 5W1H. Divide it into the items 'When, Where, Who, What, Why, How' and output it in a table format." Preferably, the instruction is set so that the desired output can be obtained. The LLM system 100 executes processing using a large-scale language model and outputs a response corresponding to prompt P2. The LLM system 100 classifies the false information from the perspective of 5W1H according to the instruction. The LLM system 100 outputs the false information classified from the perspective of 5W1H. Prompt P2 shown in Figure 3 is an example. For example, prompts could be separated for each item representing the 5W1H perspective. For example, prompts could be separated for each piece of misinformation.

[0028] Figure 4 is a table showing an example of information classified by the classification unit of the information analysis device in this disclosure. For example, the information classified in table 130 in Figure 4 was output in tabular format from the LLM system 100 in response to prompt P2 in Figure 3. Table 130 includes false information classified from the perspective of the 5W1H: "when, where, who, what, why, and how." For example, the first piece of false information was posted on SNS_Y on September 21, 2024, by account name ABC. The first piece of false information is an example of a post that criticized Country A, stating that "fuel oil from airplanes and other sources has leaked from a military base in Country A, causing severe environmental pollution in rivers and the sea," and was accompanied by images.

[0029] The evaluation unit 15 evaluates the relationships between pieces of misinformation. These relationships are reflected in the information sender, the timing of the information's transmission, and the content of the transmitted information.

[0030] For example, the evaluation unit 15 evaluates the degree of relevance between pieces of false information based on the relationship of the senders of the false information. The degree of relevance between pieces of false information sent by the same sender is high. Also, false information is often spread organizationally. Therefore, the degree of relevance between pieces of false information sent by senders with the same nationality or affiliated organization is high. On the other hand, it is presumed that the degree of relevance between pieces of false information sent by senders with completely different nationalities or affiliated organizations is low. Note that even if the senders have completely different nationalities or affiliated organizations, there may be some connection. Therefore, it is preferable to evaluate the degree of relevance between pieces of false information based on relationships other than nationality or affiliated organization.

[0031] For example, the evaluation unit 15 evaluates the degree of relevance between pieces of false information based on the timing when the false information was sent. False information is often sent in accordance with the holding of social events. Therefore, the degree of relevance between pieces of false information whose sending timings are close is high. For example, false information sent in accordance with the timing of an international event is likely to include false information directed at a specific country.

[0032] For example, the evaluation unit 15 evaluates the degree of relevance of false information based on the content of the false information sent. The words included in false information have characteristics. Therefore, the degree of relevance between pieces of false information that share the words included in the false information is high.

[0033] Also, the evaluation unit 15 evaluates the influence degree of false information. The influence degree of false information is quantified by an index representing the reaction to the false information. For example, the evaluation unit 15 is configured to evaluate the influence degree of false information by calculating a quantified index (influence degree score). For example, the evaluation unit 15 quantifies the influence degree of false information based on indicators such as a reach index, an engagement index, an audience growth index, a content analysis index, etc.

[0034] Reach metrics include reactions, reach, impressions, and hashtag usage frequency. Reactions are the total number of reactions (likes, comments, shares, etc.) to a post. Reach is the total number of unique users who saw the post. Impressions are the total number of times the post was displayed. Impressions are counted even if the same user viewed the post multiple times. Hashtag usage frequency is the number of times a particular hashtag was used.

[0035] Engagement metrics include engagement rate, share rate, click-through rate, and mentions. The engagement rate is calculated by dividing the total number of reactions to a post (likes, comments, shares, etc.) by the number of followers. The share rate is calculated by dividing the number of times a post was shared by the number of followers. The click-through rate is calculated by dividing the number of times links in a post were clicked (clicks) by the number of times the post was viewed (impressions). The number of mentions is the number of times the poster was mentioned by other users.

[0036] Audience growth metrics include follower growth rate and virality coefficient. Follower growth rate is the rate of increase in the number of followers over a certain period. It is calculated by subtracting the number of followers lost from the number of new followers and dividing by the total number of followers. Virality coefficient is the number of new users generated from a single post. It is calculated by dividing the number of new users by the number of existing users.

[0037] Content analysis metrics include sentiment analysis scores and posting frequency. The sentiment analysis score is a numerical indicator that quantifies the emotional tendencies of posts and comments. Posts and comments contain words that indicate emotional tendencies such as positive, negative, and neutral. By systematizing these emotional tendencies beforehand, the positive / negative and magnitude of these words can be quantified. Posting frequency indicates the number of posts made within a certain period.

[0038] For example, the evaluation unit 15 may be configured to evaluate the impact of misinformation by combining multiple indicators to calculate an overall impact score. For example, the evaluation unit 15 may be configured to evaluate the impact of misinformation based on the pattern of misinformation dissemination.

[0039] Figure 5 is a table showing an example of the evaluation results by the evaluation unit of the information analysis device in this disclosure. Table 150 in Figure 5 is related to Table 130 in Figure 4. Table 150 includes the relationships between misinformation, the number of reactions to the misinformation, and the degree of impact. For example, the first and second pieces of misinformation are related in that the "What" item concerns the environment in region B. For example, the first and fourth pieces of misinformation are related in that the "Why" item concerns criticism of country A. For example, the first and fifth pieces of misinformation are related in that the account names in the "Who" item are the same. In Table 150, the indicator showing the degree of impact (impact score) is the number of reactions. The degree of impact is classified by the magnitude of the impact score. A threshold for classifying the degree of impact is set in advance. In Table 150, the degree of impact of the first, second, and fourth pieces of misinformation is "medium" according to the number of reactions. Furthermore, in Table 150, the impact of the third and fifth pieces of misinformation is "high" depending on the number of reactions.

[0040] The analysis unit 16 analyzes the narrative of the disinformation sender using the evaluation results from the evaluation unit 15 to estimate the sender's true intentions and actions. In other words, the analysis unit 16 analyzes the narrative of the disinformation sender to estimate the true intentions and actions of the entity (mastermind) such as a state or organization manipulating the sender from behind the scenes. The analysis unit 16 analyzes the narrative of the disinformation sender based on background information used to predict the sender's narrative. For example, the background information used to predict the sender's narrative includes information obtained through trend analysis for each entity such as a specific person, organization, or state. The background information used to predict the sender's narrative can be stored in advance in a memory unit (not shown). The analysis unit 16 may be configured to refer to background information stored in an external database (not shown).

[0041] For example, the analysis unit 16 analyzes the narrative of a disinformation sender by referring to parameters that indicate background information for predicting the sender's narrative. The sender's intended purpose is revealed in the timing of the disinformation's dissemination. For example, disinformation is often disseminated to discredit a country before elections, before and after military exercises, or before and after national events such as signs of an impending military invasion. Disinformation disseminated at such times may contain parameters that indicate background information. Furthermore, the sender's intended purpose is revealed in the content of the disinformation message. For example, disinformation messages often contain content that suggests the victim is at fault, or content that tarnishes the reputation of allies or allies. Disinformation containing such messages may contain parameters that indicate background information.

[0042] Figure 6 is a table showing an example of parameters referenced by the analysis unit of the information analysis device in this disclosure. Table 160 includes parameters when the target country is Country C. For example, the parameter "Policy issues in Country D" is associated with specific examples such as constitutional amendment, power plants, the right of self-defense, Shinto shrine visits, island disputes, moral education, welfare, childcare, separate surnames for married couples, and same-sex marriage. In other words, disinformation containing topics such as constitutional amendment, power plants, the right of self-defense, Shinto shrine visits, island disputes, moral education, welfare, childcare, separate surnames for married couples, and same-sex marriage can predict a narrative regarding the parameter "Policy issues in Country D". For example, the parameter "Opinion leaders on issues" is associated with specific examples such as heads of state and influential politicians. In other words, disinformation containing specific examples such as heads of state and influential politicians can predict a narrative regarding the parameter "Opinion leaders on issues". Disinformation containing other specific examples can also predict a narrative regarding parameters associated with those specific examples.

[0043] For example, the analysis unit 16 classifies the extracted misinformation via an LLM system that performs processing using a large-scale language model. For example, the analysis unit 16 analyzes the narrative of the disseminator of the misinformation to estimate the original purpose and actions intended by the disseminator. The disseminator includes the mastermind who instructed the person who disseminated the misinformation to do so.

[0044] Figure 7 is a conceptual diagram showing an example of information analysis by the analysis unit of the information analysis device in this disclosure. In the example in Figure 7, the analysis unit 16 is connected to the LLM system 100 via API. The analysis unit 16 sends prompt P3 to the LLM system 100 via API. Prompt P3 is input to the LLM system 100. Prompt P3 includes instructions and the content of the false information. The instructions include the text information: "Refer to the specific examples for each parameter below, analyze the narrative of the mastermind corresponding to the source of the false information, and estimate the original purpose and actions intended by the sender." Preferably, the instructions are set so that a desired output can be obtained. The LLM system 100 performs processing using a large-scale language model and outputs a response corresponding to prompt P3. The LLM system 100 outputs an analysis result including the original purpose and actions intended by the sender, according to the instructions. Prompt P3 shown in Figure 7 is an example. For example, there may be separate prompts for setting parameters and for analyzing false information.

[0045] The output unit 17 is connected to a terminal device (not shown) used by an operator who analyzes false information. The output unit 17 obtains analysis results regarding the narrative of the false information sender from the analysis unit 16. The output unit 17 outputs analysis data, including the obtained analysis results, to the terminal device. The analysis results included in the analysis data output to the terminal device are displayed on the screen of that terminal device.

[0046] The analysis data may include advice based on the analysis results. This advice may include information to widely inform the public that disseminated information is misinformation. For example, the advice may be generated based on predetermined criteria. For example, the advice may be generated using the LLM system 100.

[0047] Figure 8 is a conceptual diagram showing an example of the display of information output from the output unit of the information analysis device in this disclosure. The screen of the terminal device 180 used by the operator displays the analysis results included in the analysis data. The analysis results include the information, "Disinformation 1, Disinformation 2, Disinformation 4, and Disinformation 5 are presumed to be narratives that will benefit militarily rising country C." By viewing the analysis results displayed on the screen of the terminal device 180, the operator can correctly understand the original purpose and actions intended by the disseminators of the disinformation. In addition, the screen of the terminal device 180 displays advice corresponding to the analysis results. The advice includes a recommendation for action, "We recommend that you widely disseminate information warning that content posted on social media from 9 / 21 to 23 related to the base of the A country's military in region B will benefit militarily rising country C." By viewing the advice displayed on the screen of the terminal device 180, the operator can take action in accordance with that advice. For example, the operator can widely notify the public of the disinformation spreading on social media through government agencies or media organizations. For example, an operator can contact the operator of a social networking service where misinformation is spreading and request that they take action against the spreading misinformation. For example, an operator can request the operator of a social networking service to freeze the account of the person spreading the misinformation or to delete the misinformation.

[0048] (Operation) Next, an example of the operation of the information analysis device in this disclosure will be described with reference to the drawings. The following description of the operation (information analysis method) is general in nature. Details of the following operation (information analysis method) are as described in the configuration described above.

[0049] Figure 9 is a flowchart illustrating an example of the operation of the information analysis device in this disclosure. In describing the process according to the flowchart in Figure 9, the components of the information analysis device 10 are considered the main operating entities. The main operating entities for the process according to the flowchart in Figure 9 may also be the information analysis device 10. For example, the process according to the flowchart in Figure 9 is realized by a processor executing a program stored in the memory of a computer (not shown) on which the information analysis device 10 is implemented.

[0050] In Figure 9, first, the collection unit 11 collects information that has been disseminated on social media (step S11).

[0051] Next, the extraction unit 12 extracts false information from the collected information (step S12).

[0052] Next, the classification unit 13 classifies the extracted false information (step S13).

[0053] Next, the evaluation unit 15 evaluates the relationships and influences between the false information (step S14).

[0054] Next, the analysis unit 16 analyzes the narrative of the disseminator of the false information (step S15).

[0055] Next, the output unit 17 outputs information corresponding to the narrative analysis results (step S16).

[0056] (Diffusion Patterns) Next, we will explain some examples of the diffusion patterns of information that are the target of analysis by the information analysis device of this embodiment. In the following, we will illustrate the diffusion patterns of false information and the diffusion patterns of misinformation. The following information diffusion patterns are just examples and do not limit the diffusion patterns of information that are the target of analysis by the information analysis device of this embodiment.

[0057] [Diffusion Pattern 1] Figures 10 to 13 are conceptual diagrams illustrating an example of the diffusion pattern of information that is the target of analysis by the information analysis device in this disclosure. Diffusion Pattern 1 is an example in which an attacker spreads false information.

[0058] Figure 10 is a conceptual diagram illustrating an example of the information dissemination pattern of the information analyzed by the information analysis device in this disclosure. The attacker disseminates false information I through a specific social networking service (SNS), which is a type of social media. d It sends out. Among the multiple users that make up the SNS user group, there is false information I sent out by the attacker. d They may react in some way to it. Misinformation I that receives a reaction d This spreads through various interpretations by multiple users who make up a group of SNS users. (Disinformation I) d This can change during the process of dissemination. The information analysis device 10 analyzes the false information I that has been disseminated on SNS. d Collect and analyze.

[0059] Figure 11 is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device described in this disclosure. Figure 11 shows an example of a response based on the analysis results of false information. In the example in Figure 11, based on the analysis results of the false information, relevant organizations are notified that false information is being spread on social media. For example, relevant organizations may be government agencies or media organizations. For example, government agencies or media organizations can publish the false information spreading on social media and warn people not to be deceived by that false information.

[0060] Figure 12 is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. Figure 12 shows an example of a response based on the analysis results of false information. In the example in Figure 12, based on the analysis results of the false information, the SNS operator is notified that false information is being spread on the SNS. The SNS operator can then issue a warning to users about the false information spreading on their SNS. The SNS operator can also issue a warning to the attacker who is the source of the false information. If the attacker has personally disseminated false information, they can delete the false information they have disseminated or take action against the spread of the false information by following the warning from the SNS operator. If the attacker has systematically disseminated false information, the warning from the SNS operator may lead to a suppression of the dissemination of false information to the entire organization.

[0061] Figure 13 is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. Figure 13 shows an example of a response based on the analysis results of false information. In the example in Figure 13, based on the analysis results of the false information, the SNS operator is notified that false information is being spread on the SNS. The SNS operator can then send a warning to users about the false information spreading on the SNS. The SNS operator can also freeze or delete the account of the attacker who is the source of the false information. Furthermore, the SNS operator can make it known on the SNS that the false information is malicious and send correct information related to the false information. c By spreading this information, it is possible to purify the information that has been disseminated on that social media platform.

[0062] [Diffusion Pattern 2] Figures 14 to 16 are conceptual diagrams illustrating an example of an information diffusion pattern that is the target of analysis by the information analysis device in this disclosure. Diffusion Pattern 2 is an example in which information disseminated on social networking services (SNS) is altered and disseminated by an attacker.

[0063] FIG. 14 is a conceptual diagram for explaining an example of a diffusion pattern of information to be analyzed by the information analysis apparatus in the present disclosure. An information sender transmits correct information I via a specific SNS which is one of social media. c Among a plurality of users constituting the SNS user group, some may react to the correct information I transmitted by the information sender. c The correct information I that has received a reaction is diffused while being subject to various interpretations by a plurality of users constituting the SNS user group. An attacker tampers with the diffused correct information I into false information I. c The tampered false information I is diffused while being subject to various interpretations by a plurality of users constituting the SNS user group. The correct information I and the false information I may change during the diffusion process. The information analysis apparatus 10 collects and analyzes the correct information I and the false information I diffused in the SNS. c d d c d c d

[0064] FIG. 15 is a conceptual diagram for explaining an example of a response according to the analysis result by the information analysis apparatus in the present disclosure. FIG. 15 shows an example of a response according to the analysis result of false information. In the example of FIG. 15, according to the analysis result of the false information, the operator of the SNS is notified that the false information is being diffused in the SNS. The operator of the SNS can transmit a message to the users to pay attention to the false information being diffused in the SNS. Further, the operator of the SNS can issue a warning to the user who diffused the false information. The users who diffused the false information include attackers and normal users. When an attacker personally transmits false information, by following the warning from the operator of the SNS, the attacker can delete the transmitted false information or take measures against the diffused false information. When an attacker diffusely transmits false information organizationally, the warning from the operator of the SNS may lead to suppression of the transmission of false information to the entire organization. Also, for normal users who diffused false information without knowing it, it is possible to prompt them to pay attention regarding information diffusion.​​​​​​​

[0065] Figure 16 is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. Figure 16 shows an example of a response based on the analysis results of false information. In the example in Figure 16, based on the analysis results of the false information, the SNS operator is notified that false information is being spread on the SNS. The SNS operator can then send a warning to users about the false information spreading on the SNS. The SNS operator can also freeze or delete the account of the attacker who is the source of the false information. Furthermore, the SNS operator can make it known on the SNS that the false information is malicious and send correct information related to the false information. c By spreading this information, it is possible to purify the information that has been disseminated on that social media platform.

[0066] [Diffusion Pattern 3] Figures 17 to 19 are conceptual diagrams illustrating an example of an information diffusion pattern that is the subject of analysis by the information analysis device in this disclosure. Diffusion Pattern 3 is an example in which information disseminated on social networking services (SNS) is misinterpreted by a typical user who is a misposter and then disseminated as misinformation.

[0067] Figure 17 is a conceptual diagram illustrating an example of the information diffusion pattern that is the target of analysis by the information analysis device in this disclosure. Information senders transmit correct information via a specific SNS, which is one of the social media platforms. c It transmits information. Among the multiple users that make up the SNS user group, there is correct information transmitted by the information sender. c They may react in some way to it. Correct information received from the reaction I c This information spreads through various interpretations by multiple users who make up a group of SNS users. The person who spread the misinformation is the one who disseminated the correct information. c Misinterpreting this leads to misinformation I m This is being disseminated. Misinformation I m This information spreads through various interpretations by multiple users who make up the SNS user base. c and misinformation I mThis information may change during the process of dissemination. The information analysis device 10 analyzes the correct information I that has been disseminated on SNS. c and misinformation I m Collect and analyze the information. Note: Misinformation I m This information is extracted by the information analysis device 10 in the same way as false information, and is determined to be misinformation by analyzing its narrative.

[0068] Figure 18 is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. Figure 18 shows an example of a response based on the analysis results of misinformation. In the example in Figure 18, based on the analysis results of the misinformation, the SNS operator is notified that misinformation is being spread on the SNS. The SNS operator can then send a warning to users about the misinformation being spread on the SNS. The SNS operator can also issue a warning to the user who was the source of the misinformation (the misinformation sender). Users who spread misinformation include both the misinformation sender and regular users. The misinformation sender can delete the misinformation they sent or take action against the spread of the misinformation by following the warning from the SNS operator. Regular users who unknowingly spread misinformation can also be warned about the spread of information.

[0069] Figure 19 is a conceptual diagram illustrating an example of a response based on the analysis results of the information analysis device in this disclosure. Figure 19 shows an example of a response based on the analysis results of misinformation. In the example in Figure 19, based on the analysis results of the misinformation, the SNS operator is notified that misinformation is being spread on the SNS. The SNS operator can then send a warning to users about the misinformation being spread on the SNS. The SNS operator can also recommend that the user who is the source of the misinformation (the misinformation sender) take literacy training on information dissemination. For example, the SNS operator can take measures such as temporarily freezing the user's account until the user who is the misinformation sender completes the literacy training. Furthermore, the SNS operator can make it known that the misinformation spread on the SNS is false and send correct information related to the misinformation. c By spreading this information, it is possible to purify the information that has been disseminated on that social media platform.

[0070] As described above, the information analysis device of this embodiment comprises a collection unit, an extraction unit, a classification unit, an evaluation unit, an analysis unit, and an output unit. The collection unit collects information disseminated on social media. The extraction unit extracts false information from the collected information in accordance with fact-checking standards. The classification unit classifies the extracted false information. The evaluation unit evaluates the relevance and impact of the classified false information. The analysis unit analyzes the narrative of the false information sender by referring to parameters that indicate background information for predicting the sender's narrative and the relevance and impact of the false information. The output unit outputs analysis data including the analysis results of the false information sender's narrative.

[0071] The information analysis device of this embodiment evaluates the relevance and impact of misinformation extracted in accordance with fact-checking standards and analyzes the narrative of the disseminator of the misinformation. By analyzing the narrative of the disseminator of the misinformation, the information analysis device of this embodiment can accurately interpret the disseminator's purpose and intentions. Therefore, according to this embodiment, analysis data including analysis results that accurately reflect the dissemination of misinformation's purpose and intentions makes it possible to correctly interpret that disseminated misinformation is false.

[0072] In one embodiment of this system, the collection unit collects false information disseminated on social media. The analysis unit analyzes the narrative of the disseminator of the collected false information. According to this embodiment, the process of extracting false information from information disseminated on social media can be omitted.

[0073] In one aspect of this embodiment, the classification unit classifies false information from the perspective of the 5W1H. The 5W1H perspective includes the timing of when the false information was transmitted, the social networking service on which the false information was transmitted, the entity that transmitted the false information, the content contained in the false information, the purpose for which the false information was transmitted, and the form of the false information. According to this embodiment, by classifying false information from the perspective of the 5W1H, the narrative of the sender of the false information can be interpreted more accurately.

[0074] In one embodiment of this system, the evaluation unit evaluates the relevance of the false information based on at least one of the following criteria: the sender of the false information, the timing of the false information's transmission, and the content of the false information. According to this embodiment, by evaluating the relevance based on clear criteria for the false information, the narrative of the sender of the false information can be interpreted more accurately.

[0075] In one embodiment of this system, the evaluation unit assesses the impact of misinformation using an impact score quantified by an index representing the reaction to the misinformation. According to this embodiment, by evaluating the quantified impact score, the narrative of the person who disseminated the misinformation can be interpreted more accurately.

[0076] (Second Embodiment) Next, the information analysis device in the second embodiment will be described with reference to the drawings. The information analysis device in this embodiment has a simplified configuration compared to the information analysis device in the first embodiment. For example, the functions of the components of the information analysis device in this embodiment are realized by the functions of the components of the information analysis device according to the first embodiment.

[0077] (Configuration) Figure 20 is a block diagram showing an example of the configuration of an information analysis device in this disclosure. The information analysis device 20 comprises a collection unit 21, an analysis unit 26, and an output unit 27.

[0078] The collection unit 21 collects information disseminated on social media. The analysis unit 26 analyzes the narrative of the information sender. The output unit 27 outputs analysis data including the results of the analysis of the information sender's narrative.

[0079] (Operation) Figure 21 is a flowchart showing an example of the operation of the information analysis device in this disclosure. In the explanation of the process according to the flowchart in Figure 21, the components of the information analysis device 20 are considered the main operating components. The main operating component of the process according to the flowchart in Figure 21 may also be the information analysis device 20.

[0080] In Figure 21, first, the collection unit 21 collects information that has been disseminated on social media (step S21).

[0081] Next, the analysis unit 26 analyzes the narrative of the information sender (step S22).

[0082] Next, the output unit 27 outputs analysis data including the results of the analysis of the information sender's narrative (step S23).

[0083] In this embodiment, by analyzing the narrative of the sender of information disseminated on social media, it is possible to accurately interpret the sender's purpose and intentions. Therefore, according to this embodiment, it becomes possible to enable people to correctly interpret the disseminated information through analytical data that includes analysis results that accurately reflect the sender's purpose and intentions.

[0084] (Hardware) Next, the hardware configuration for performing the processing described in this disclosure will be described with reference to the drawings. Figure 22 is a block diagram showing an example of a hardware configuration for performing the processing described in this disclosure. Here, an information processing device 90 (computer) is shown as an example of a hardware configuration. The information processing device in Figure 22 is an example configuration for performing the processing described in this disclosure and does not limit the scope of this disclosure.

[0085] As shown in Figure 22, the information processing device 90 includes a processor 91, memory 92, auxiliary storage device 93, input / output interface 95, and communication interface 96. In Figure 22, interface is abbreviated as I / F (Interface). The information processing device 90 may include at least one or more of the processor 91, memory 92, auxiliary storage device 93, input / output interface 95, and communication interface 96. The processor 91, memory 92, auxiliary storage device 93, input / output interface 95, and communication interface 96 are connected to each other via a bus 98 so that they can communicate data. In addition, the processor 91, memory 92, auxiliary storage device 93, and input / output interface 95 are connected to a network such as the Internet or an intranet via the communication interface 96.

[0086] The processor 91 loads a program (instructions) stored in an auxiliary storage device 93 or the like into memory 92. For example, the program is a software program for executing the processing described in this disclosure. The processor 91 executes the program loaded into memory 92. The processor 91 executes the processing described in this disclosure by executing the program. The processor 91 may be composed of a single piece of hardware or of multiple pieces of hardware.

[0087] Memory 92 is a storage device having an area where a program is loaded. The processor 91 loads the program stored in the auxiliary storage device 93 or the like into memory 92. Memory 92 can be implemented using volatile memory such as DRAM (Dynamic Random Access Memory). Alternatively, non-volatile memory such as MRAM (Magnetoresistive Random Access Memory) may be used as memory 92. Memory 92 may be composed of a single piece of hardware or multiple pieces of hardware.

[0088] The auxiliary storage device 93 stores various data, such as programs. For example, the auxiliary storage device 93 can be implemented by a local disk such as a hard disk or flash memory. The auxiliary storage device 93 may be configured by a single piece of hardware or by multiple pieces of hardware. The auxiliary storage device 93 may also be configured as external hardware. It is also possible to configure the system to store various data in memory 92 and omit the auxiliary storage device 93.

[0089] The input / output interface 95 is an interface for connecting the information processing device 90 to peripheral devices based on standards and specifications. The communication interface 96 is an interface for connecting to external systems and devices via a network such as the Internet or an intranet, based on standards and specifications. The input / output interface 95 may be composed of a single piece of hardware or multiple pieces of hardware. The input / output interface 95 and the communication interface 96 may be common as interfaces for connecting to external devices.

[0090] The information processing device 90 may be connected to input devices such as a keyboard, mouse, or touch panel, as needed. These input devices are used to input information and settings. When a touch panel is used as an input device, the screen with touch panel functionality becomes the interface. The processor 91 and the input devices are connected via an input / output interface 95.

[0091] The information processing device 90 may be equipped with a display device for displaying information. If a display device is provided, the information processing device 90 is equipped with a display control device (not shown) for controlling the display of the display device. The information processing device 90 and the display device are connected via an input / output interface 95.

[0092] The information processing device 90 may be equipped with a drive device. The drive device mediates between the processor 91 and the recording medium (program recording medium) by reading data and programs stored on the recording medium and writing the processing results of the information processing device 90 to the recording medium. The information processing device 90 and the drive device are connected via an input / output interface 95.

[0093] The above is an example of a hardware configuration to enable the processing described in this disclosure. The hardware configuration in Figure 22 is an example of a hardware configuration for executing the processing described in this disclosure and does not limit the scope of this disclosure. A program that causes a computer to execute the processing described in this disclosure is also included in the scope of this disclosure.

[0094] A program recording medium that stores a program for performing the processing in this embodiment is also included in the scope of this disclosure. For example, the program recording medium is a computer-readable, non-transient recording medium. The recording medium can be implemented as an optical recording medium such as a CD (Compact Disc) or a DVD (Digital Versatile Disc). The recording medium may also be implemented as a semiconductor recording medium such as a USB (Universal Serial Bus) memory or an SD (Secure Digital) card. Furthermore, the recording medium may be implemented as a magnetic recording medium such as a flexible disk, or other recording media.

[0095] The components in this disclosure may be combined in any way. The components in this disclosure may be implemented by software. The components in this disclosure may be implemented by circuitry.

[0096] Although the present disclosure has been described above with reference to embodiments, the present disclosure is not limited to the embodiments described above. Various modifications to the structure and details of the present disclosure can be made as can be understood by those skilled in the art within the scope of the present disclosure. Furthermore, each embodiment can be combined with other embodiments as appropriate.

[0097] Some or all of the above embodiments may also be described as follows, but are not limited to the following. In the following appendices, the dependent terms of each category may also be dependent on other categories. The descriptions included in the following appendices have significance as grounds for amendment. (Appendix 1) An information analysis device comprising: a collection unit that collects information disseminated on social media; an analysis unit that analyzes the narrative of the sender of the collected information; and an output unit that outputs analysis data including the analysis results of the narrative of the sender of the information. (Appendix 2) The information analysis device according to Appendix 1, wherein the analysis unit analyzes the narrative of the sender of the information by referring to parameters that indicate background information for predicting the sender's narrative. (Appendix 3) The information analysis device according to Appendix 2, wherein the collection unit collects false information disseminated on social media, and the analysis unit analyzes the narrative of the sender of the collected false information. (Note 4) An information analysis device according to Note 2, comprising an extraction unit that extracts false information from the collected information in accordance with the norms of fact-checking, wherein the analysis unit analyzes the narrative of the sender of the extracted false information. (Note 5) An information analysis device according to Note 3 or 4, comprising a classification unit that classifies the extracted false information, and an evaluation unit that evaluates the relevance and impact of the classified false information, wherein the analysis unit analyzes the narrative of the sender of the false information by referring to the relevance and impact of the false information. (Note 6) An information analysis device according to Note 5, wherein the classification unit classifies the false information in terms of the 5W1H, including the timing of when the false information was transmitted, the social networking service on which the false information was transmitted, the entity that transmitted the false information, the content contained in the false information, the purpose for which the false information was transmitted, and the format of the false information. (Note 7) An information analysis device according to Note 5, wherein the evaluation unit evaluates the relevance of the false information based on at least one of the sender of the false information, the timing of when the false information was transmitted, and the content of the false information. (Note 8) The evaluation unit is an information analysis device according to Note 5 that evaluates the degree of influence of the false information by an influence score quantified by an index representing the response to the false information.(Note 9) An information analysis method in which a computer collects information disseminated on social media, analyzes the narrative of the information sender that was collected, and outputs analysis data including the results of the analysis of the information sender's narrative. (Note 10) A program that causes a computer to perform the following: a process of collecting information disseminated on social media, a process of analyzing the narrative of the information sender that was collected, and a process of outputting analysis data including the results of the analysis of the information sender's narrative. Furthermore, some or all of the configurations described in Notes 2 to 8, which are dependent on Note 1 above, may also be dependent on Notes 9 and 10 in the same way as in Notes 2 to 8. Moreover, not limited to Notes 1, 9, and 10, some or all of the configurations described as notes may also be dependent on various hardware, software, various recording means for recording software, or systems, without departing from the embodiments described above.

[0098] This application claims priority based on Japanese Patent Application No. 2024-224821, filed on 20 December 2024, and incorporates all of its disclosures herein.

[0099] 10, 20 Information analysis device 11, 21 Collection unit 12 Extraction unit 13 Classification unit 15 Evaluation unit 16, 26 Analysis unit 17, 27 Output unit

Claims

1. An information analysis device comprising: a collection unit that collects information disseminated on social media; an analysis unit that analyzes the narrative of the information sender; and an output unit that outputs analysis data including the analysis results of the information sender's narrative.

2. The information analysis device according to claim 1, wherein the analysis unit analyzes the narrative of the information sender by referring to parameters that indicate background information for predicting the sender's narrative.

3. The information analysis device according to claim 2, wherein the collection unit collects false information disseminated on social media, and the analysis unit analyzes the narrative of the sender of the collected false information.

4. The information analysis device according to claim 2, comprising an extraction unit that extracts false information from the collected information in accordance with fact-checking standards, wherein the analysis unit analyzes the narrative of the source of the extracted false information.

5. The information analysis device according to claim 3 or 4, comprising: a classification unit for classifying the extracted false information; and an evaluation unit for evaluating the relevance and influence of the classified false information, wherein the analysis unit analyzes the narrative of the sender of the false information by referring to the relevance and influence of the false information.

6. The information analysis device according to claim 5, wherein the classification unit classifies the false information from the perspective of the 5W1H, including the timing of when the false information was transmitted, the social networking service on which the false information was transmitted, the entity that transmitted the false information, the content contained in the false information, the purpose for which the false information was transmitted, and the format of the false information.

7. The information analysis device according to claim 5, wherein the evaluation unit evaluates the degree of relevance of the false information based on at least one of the sender of the false information, the timing at which the false information was transmitted, and the content of the false information.

8. The information analysis device according to claim 5, wherein the evaluation unit evaluates the degree of influence of the false information using an influence score quantified by an index representing the response to the false information.

9. An information analysis method comprising a computer that collects information disseminated on social media, analyzes the narrative of the information sender, and outputs analytical data including the results of the analysis of the information sender's narrative.

10. A program that causes a computer to perform the following processes: collecting information disseminated on social media; analyzing the narrative of the information sender; and outputting analytical data including the results of the analysis of the information sender's narrative.

Images