Evaluation device, evaluation program, and evaluation method

Abstract

Provided is an evaluation device that evaluates material information constituting a material for advertisement, the evaluation device comprising: an advertisement rule determination unit that determines, on the basis of the material information, an advertisement rule related to an advertisement in which the material information is used; and an evaluation unit that evaluates the material information on the basis of the advertisement rule.

Description

Evaluation device, evaluation program, and evaluation method

[0001] The present invention relates to an evaluation device, an evaluation program, and an evaluation method.

[0002] A technology has been disclosed that evaluates the performance of advertisements and provides suggestions for making the content of advertisements more effective.

[0003] JP 2011-008792 A

[0004] Patent Literature 1 discloses a method for evaluating the performance of web advertisements and adjusting the content related to branding. However, this technology cannot evaluate whether the content of the advertisements violates various rules, such as laws.

[0005] The present invention has been made in view of the above background, and aims to support the creation of advertisements that do not pose any problems in terms of various rules.

[0006] In order to solve the above problems, the evaluation device for evaluating advertising materials in the present disclosure includes an advertising rule determination unit that determines advertising rules related to advertisements in which the material information is used based on the material information, and an evaluation unit that evaluates the material based on the advertising rules.

[0007] Other problems and solutions disclosed in this application will be made clear in the section on preferred embodiments of the invention and the drawings.

[0008] According to the present invention, it is possible to support the creation of advertisements that do not pose any problems in compliance with various rules.

[0009] FIG. 1 is a diagram showing an example of the overall configuration of an evaluation system according to an embodiment of the present invention. FIG. 2 is a diagram showing an example of the hardware configuration of a server device 1 according to the embodiment. FIG. 3 is a diagram showing an example of the functional configuration of the server device 1 according to the embodiment. FIG. 4 is a diagram showing an example of actual consumption information stored in a material information storage unit 131 according to the embodiment. FIG. 5 is a diagram showing an example of inventory transition schedule information stored in an advertising rule information storage unit 132 according to the embodiment. FIG. 6 is a diagram showing an example of edited material information stored in a edited material information storage unit 133 according to the embodiment. FIG. 7 is a diagram showing an example of processing of the server device 1 according to the embodiment.

[0010] <Summary of the Invention> The details of embodiments of the present invention will be listed below. The present invention, for example, has the following configuration. [Item 1] An evaluation device for evaluating material information, which is material for advertising, comprising: an advertising rule determination unit that determines advertising rules related to advertisements in which the material information is used, based on the material information; and an evaluation unit that evaluates the material information based on the advertising rules. [Item 2] The evaluation device according to Item 1, wherein the material includes text information and image information, and the advertising rule determination unit analyzes the text information to determine the advertising rules. [Item 3] The evaluation device according to Item 1 or 2, wherein the evaluation unit evaluates the material information based on the advertising rules. [Item 4] The evaluation device according to Item 3, further comprising: an edited material information generation unit that generates non-problematic edited material information when the evaluation unit evaluates the material information as problematic under the advertising rules. [Item 5] The evaluation device according to Item 4, further comprising: a presentation unit that presents the problematic material information and the edited material information to a user based on the evaluation result of the evaluation unit. [Item 6] An evaluation program for evaluating material information for advertising, the evaluation program causing a computer to execute: an advertising rule determination step of determining, based on the material, advertising rules associated with an advertisement in which the material information is used; and an evaluation step of evaluating the material based on the advertising rules. [Item 7] An evaluation method for evaluating material information for advertising, the evaluation method causing a computer to execute: an advertising rule determination step of determining, based on the material, advertising rules associated with an advertisement in which the material information is used; and an evaluation step of evaluating the material based on the advertising rules. <Details of embodiment>

[0011] Preferred embodiments of the present disclosure will be described in detail below with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0012] ==Overview== Figure 1 is a diagram showing the overall configuration of the evaluation system. As shown in Figure 1, the evaluation system has a server device 1 and a user terminal 3. The server device 1 is connected to the user terminal 3 via a network 2. Although only one server device 1 and one user terminal 3 are shown, it goes without saying that there may be more than one server device 1 and more than one user terminal 3.

[0013] ==Server Device 1== The server device 1 may be, for example, a general-purpose computer such as a workstation or a personal computer, or may be logically realized by cloud computing.

[0014] ==User Terminal 3== The user terminal 3 is a computer operated by a user who evaluates advertisements. For example, it is a smartphone, a tablet computer, a personal computer, etc. The user can access the server device 1 using, for example, an application or a web browser executed on the user terminal 3.

[0015] FIG. 2 is a diagram illustrating an example of the hardware configuration of the server device 1. Note that the illustrated configuration is an example, and other configurations may also be used. The server device 1 includes a CPU 101, a memory 102, a storage device 103, a communication interface 104, an input device 105, and an output device 106. The storage device 103 stores various data and programs, and is, for example, a hard disk drive, a solid-state drive, or a flash memory. The communication interface 104 is an interface for connecting to the communication network 2, and is, for example, an adapter for connecting to Ethernet (registered trademark), a modem for connecting to a public telephone network, a wireless communication device for wireless communication, a USB (Universal Serial Bus) connector, or an RS232C connector for serial communication. The input device 105 is, for example, a keyboard, a mouse, a touch panel, a button, a microphone, or the like for inputting data. The output device 106 is, for example, a display, a printer, a speaker, or the like for outputting data. Each functional unit of the server device 1 described below is realized by the CPU 101 reading a program stored in the storage device 103 into the memory 102 and executing it, and each storage unit of the server device 1 is realized as part of the storage area provided by the memory 102 and the storage device 103.

[0016] Fig. 3 also shows the functional configuration of the server device 1. As shown in Fig. 3, the server device 1 includes storage units, namely, a material information storage unit 131, an advertising rule information storage unit 132, and an edited material information storage unit 133, as well as processing units, namely, a material information acquisition unit 111, a material information analysis unit 112, an advertising rule determination unit 113, an evaluation unit 114, an edited material information generation unit 115, and a presentation unit 116.

[0017] The material information storage unit 131, the advertising rule information storage unit 132, and the edited material information storage unit 133 will be described below.

[0018] The material information storage unit 131 stores advertisement material information acquired by the material information acquisition unit 111, an example of which is shown in Fig. 4. As shown in an example in Fig. 4, the material information includes information such as text information used in the advertisement, such as a catchphrase, tagline, description, and image text information obtained by reading characters included in image information used in the advertisement, and image information (including video) used in the advertisement, and information on the target field or target product analyzed by the material information analysis unit 112 (described later) is stored in association with each piece of material information.

[0019] The advertising rule information storage unit 132 stores rules related to advertisements. As shown in an example in Fig. 5 , the advertising rule information is linked to the target field of the advertisement or the target product of the advertisement, and includes information such as related laws and regulations, regulations of industry organizations, internal rules and guidelines of users, and what readers think (emotional evaluation and marketing elements), as well as information on prohibited expressions and expressions that should be avoided.

[0020] The edited material information storage unit 133 stores the edited material information generated by the edited material information generation unit 115. As shown in an example in Fig. 6, the edited material may include text information, image information, combined image information that combines the text information and the image information, and the like.

[0021] The following describes each processing unit, namely, the material information acquisition unit 111, the material information analysis unit 112, the advertising rule determination unit 113, the evaluation unit 114, the edited material information generation unit 115, and the presentation unit 116.

[0022] As an example, the material information acquisition unit 111 acquires material information for advertisements. For example, the material information acquisition unit 111 presents an input form for material information to the user terminal 3 via the network 2, accepts user input operations, and acquires the material information. The communication for sending and receiving the information may be wired or wireless, and any communication protocol may be used as long as mutual communication is possible. Note that the material information acquisition unit 111 may, for example, accept and acquire document files, image files, etc. containing advertising material using the method described above, or may acquire the URL of a website containing advertising information from the user terminal and acquire information on the website as material information, but is not limited to these methods. The material information acquisition unit 111 stores the acquired material information in the material information storage unit 131.

[0023] The material information acquisition unit 111 acquires text information (particularly referred to as image text information) from image information included in the material information, for example. The material information acquisition unit 111 acquires character information included in the image as image text information using existing image analysis technology such as OCR. The material information acquisition unit 111 stores the image text information in the material information storage unit 131 as part of text information acquired separately.

[0024] As an example, the material information analysis unit 112 analyzes the material information acquired by the material information acquisition unit 111 .

[0025] For example, the material information analysis unit 112 generates prompt information based on the material information. The prompt information may include a prompt or a prerequisite condition. The material information analysis unit 112 generates the prompt information using, for example, a language model.

[0026] The material information analysis unit 112, for example, inputs prompt information generated based on the text information contained in the material information into an existing text generation model (for example, a large-scale language model such as ChatGPT), and based on the output information, identifies basic advertising information such as the target field of the advertisement in which the material information is used and the target product that the material information targets, and stores this information in the material information storage unit 131 in association with the material information.

[0027] The material information analysis unit 112 may, for example, morphologically analyze the text information using existing morphological analysis technology to divide it into words, vectorize each word, identify basic advertising information of the advertisement in which the material information is used, and store the vector information in the material information storage unit 131 in association with the material information.

[0028] The material information analysis unit 112 may, for example, analyze the image information contained in the material information using existing image analysis technology, identify the basic advertising information of the advertisement in which the material information is used, and store the information in the material information storage unit 131 in association with the material information.

[0029] The material information analysis unit 112 may identify the basic advertisement information using a model that uses, for example, image information as input information and at least some of the basic advertisement information of the advertisement in which the material information is used as output information. The model may be generated by machine learning, deep learning, or the like, or may be a large-scale image language model, multimodal AI, or the like, but is not limited to these.

[0030] The advertising rule determination unit 113 determines advertising rules to which the material information to be evaluated may be related, based on the analysis results of the material information analysis unit 112. The advertising rule determination unit 113 may determine advertising rules to which the material to be evaluated may be related, for example, based on at least some of the basic advertising information identified by the analysis of the material information analysis unit 112. The advertising rule determination unit 113 may determine, as advertising rules, advertising rules stored in association with, for example, the presence or absence of a predetermined word identified by the material information analysis unit 112, a target field, a target product, etc. For example, if the analysis of the material information analysis unit 112 identifies the target field as the food field or the target product as a health food, the advertising rule determination unit 113 may identify, as advertising rules, the Pharmaceutical and Medical Device Act stored in association with the target field or the target product.

[0031] As an example, the advertising rule determination unit 113 may input prompt information generated based on the material information acquired by the material information acquisition unit 111 into an existing text generation model, and based on the output information, identify advertising rules that may be relevant to the advertisement in which the material information is used.

[0032] For example, the advertising rule determination unit 113 may learn information linking material information with advertising rules as training data, and identify advertising rules using a model that uses the material information as input information and advertising rules that may be related to advertisements in which the material information is used as output information. The model may be generated by machine learning, deep learning, or the like, or may be a large-scale image language model, multimodal AI, or the like, but is not limited to these.

[0033] The evaluation unit 114 evaluates the material information based on the advertising rules determined by the advertising rule determination unit 113. The evaluation unit 114 evaluates whether or not there is a problem in the material information, such as whether or not there is an expression that cannot be used in order to comply with the advertising rules or an expression that may conflict with the advertising rules. Note that the presence or absence of a problem in the evaluation performed by the evaluation unit 114 may be the result of determining whether or not there is an expression that may be considered problematic.

[0034] The evaluation unit 114 evaluates whether or not there is a problem with the material information by determining whether or not the material information contains prohibited words, prohibited expressions, etc. that are stored in association with the advertising rules, for example.

[0035] The evaluation unit 114 may, for example, use a language model to generate prompt information including at least the material information and information on the advertising rules determined by the advertising rule determination unit 113, input the prompt information into an existing text generation model, and evaluate the material information based on the output information.

[0036] The evaluation unit 114 evaluates, for example, whether the image information included in the material information contains expressions that cannot be used in order to comply with the advertising rules or whether there is a part that may conflict with the advertising rules. In this case, the evaluation unit 114 may generate prompt information that includes the material information and information on the advertising rules determined by the advertising rule determination unit 113, input the prompt information to an existing large-scale image language model, and evaluate the image information based on the output information.

[0037] For example, the evaluation unit 114 may evaluate the image information using a model in which, as training data, information linking image information that includes expressions that cannot be used to comply with advertising rules or parts that may violate advertising rules with information on the expressions that cannot be used to comply with advertising rules or parts that may violate advertising rules is used, the image information is used as input information, and the expressions that cannot be used to comply with advertising rules or parts that may violate advertising rules is used as output information. The model may be generated by machine learning, deep learning, or the like, or may be a large-scale image language model, multimodal AI, image recognition model, or the like, or may use existing image processing technology, but is not limited to these.

[0038] For example, when the material information analysis unit 112 determines that the target field is the health and healthcare field, and the advertising rule determination unit 113 determines that the Pharmaceutical and Medical Device Act is the advertising rule, if the material information contains prohibited expressions such as "Anyone can lose weight!" or "Your hair will grow dramatically!", the evaluation unit 114 evaluates that the material information may violate the advertising rules.

[0039] The edited material information generation unit 115 generates edited material information, which is a proposed edit, for material information that the evaluation unit 114 has determined may be in violation of the advertising rules. The edited material information generation unit 115 generates prompt information including the material information and advertising rule information (including information on prohibited expressions, etc.) determined by the advertising rule determination unit 113, inputs the prompt information to a large-scale language model, a large-scale image language model, etc., and generates edited material information that does not infringe on the advertising rules based on the output information and stores it in the edited material information storage unit 133.

[0040] The edited material information generation unit 115 may generate edited material information using a model in which information linking the material information and the edited material information is used as training data, the material information is used as input information, and the edited material information is used as output information. The model may be generated by machine learning, deep learning, or the like, or may be a large-scale image language model, multimodal AI, image recognition model, or the like, or may use existing image processing technology, but is not limited to these.

[0041] The presentation unit 116 presents the basic advertising information, advertising rules, evaluation results, edited material information, and the like output by the above-mentioned processing units to the user terminal 3. For example, based on the evaluation results of the evaluation unit 114, the presentation unit 116 may present material information that may conflict with the advertising rules and the edited material information generated by the edited material information generation unit 115 to the user terminal 3 as an edit proposal.

[0042] If the user accepts the suggested revision, the presentation unit 116 may transmit the revision material information to the user terminal 3.

[0043] A typical processing flow of the server device 1 of this embodiment will be described using Figure 7. The material information acquisition unit 111 acquires material information (1001). The material information analysis unit 112 analyzes the material information (1002). The advertising rule determination unit 113 determines advertising rules from the analyzed material information (1003). The evaluation unit 114 evaluates the material information based on the advertising rules (1004). The edited material information generation unit 115 generates edited material information (1005). The presentation unit 116 presents the material information that conflicts with the advertising rules and the edited material information to the user (1006).

[0044] Other examples are shown below.

[0045] For example, the server device 1 may use video information as image information. In this case, the material information analysis unit 112 may acquire frames of the video information at predetermined intervals and analyze them as image information using the method described above. The material information analysis unit 112 may also analyze frames containing text as image information using the method described above.

[0046] The server device 1 may also evaluate advertising materials using video information in the following manner: The video information includes multiple media elements such as each frame image, text information included in the image, audio information, and text information obtained by speech recognition description of the audio.

[0047] The server device 1 acquires text information from frame images constituting the video using character recognition technology such as OCR, and the evaluation unit 114 checks the text information. Furthermore, the evaluation unit 114 performs an image check on the frame images and a check of the text information acquired from the audio included in the video using voice recognition technology.

[0048] In addition to checking the video itself, the server device 1 also evaluates the accompanying information that accompanies the video information, such as the title, introduction, thumbnail (title image), hashtag, icon, user reviews, and comments, and performs an overall evaluation that combines the evaluations of the video information and the accompanying information.

[0049] Furthermore, the server device 1 may verbalize the constituent information that constitutes the advertising video, i.e., the video, frame images, audio information, etc., using multimodal AI or the like, and evaluate the text information using the evaluation method described above by the evaluation unit 114. This method enables the server device 1 to use the advertisement check mechanism while also performing evaluation based on the combination of constituent information.

[0050] Furthermore, the server device 1 may directly verbalize the entire video that includes the configuration information using multimodal AI or the like, and check the generated sentences (text information).

[0051] The evaluation unit 114 checks the sentences (text information) used in the evaluation of advertising materials using video information by using large-scale language models (LLMs), natural language processing, AI-based NG sentence learning, vector detection, keyword detection, etc., to detect violations of various laws and regulations such as the Pharmaceutical and Medical Device Act and the Specified Commercial Transactions Act.

[0052] The image checks performed by the evaluation unit 114 in the evaluation of advertising materials using video information utilize multimodal AI, image search, object detection, object classification, image processing technology, etc., and can detect violation patterns such as, for example, "Is there someone wearing a white coat?" or "Is there a person wearing a white coat advertising cosmetics?"

[0053] One example of a method for improving the accuracy of advertisement evaluations performed by the server device 1 is the adoption of a complex preprocessing and a step-by-step analysis method. For example, the server device 1 may perform person detection from advertisement material information and then perform analysis using LLM or multimodal AI. Specifically, the server device 1 first identifies people in an image using, for example, an object detection algorithm, and extracts attributes such as the type of clothing (e.g., white coat, uniform), age group, and facial expression. The server device 1 then inputs the detected person information and the original image into the multimodal AI, thereby enabling more accurate detection of expressions that may violate the Pharmaceutical and Medical Device Act, such as "a person who looks like a doctor is recommending a product."

[0054] Analyzing the continuity of movements and actions is also effective. The server device 1 tracks the actions of people in a video over time and automatically detects sequences that suggest before-and-after comparisons, such as "before using a product → after using it." For example, the server device 1 identifies scenes in which changes occur in skin condition, body shape, hair condition, etc., and evaluates the likelihood that these change patterns could be interpreted as "emphasis on effects and efficacy." After performing this type of time-series behavior detection as preprocessing, the server device 1 inputs the video along with the detected behavioral pattern information into multimodal AI, enabling highly accurate identification of regulatory violation risks that are difficult to detect through simple image analysis.

[0055] Another example is a method utilizing object relationship analysis. The server device 1 detects the positional relationships and interactions between products and other objects (e.g., fruits, measuring instruments, comparison objects, etc.) in advertising images and videos as preprocessing. For example, if a health food product and a specific fruit are always displayed together, this may be interpreted as an indirect promotion of the fruit's efficacy. The server device 1 structures such inter-object relationship information and inputs it into multimodal AI, enabling more accurate detection of complex patterns such as implicit efficacy claims.

[0056] Analyzing the consistency between text and images is also an important preprocessing step. The server device 1 analyzes the consistency between the text in an image (extracted using OCR, etc.) and the image content, and evaluates the possibility of inconsistencies or exaggerations in advance. For example, if the text "clinical trial completed" is displayed, the server device 1 verifies whether the image contains backgrounds or props that evoke a laboratory or research institute. By structuring this consistency information and inputting it into the LLM, misleading statements and context-dependent regulatory violation risks can be more accurately detected.

[0057] Emotion analysis preprocessing is also an effective technique. The server device 1 detects emotional states (such as joy, surprise, or relief) from the facial expressions and postures of people in advertisements and structures patterns of emotional change before and after product use. For example, the server device 1 may interpret patterns such as facial expressions of dissatisfaction or worry before using a product and expressions of joy or relief after use as indirect appeals to its effectiveness or efficacy. By extracting such emotional change information as preprocessing and inputting it into the LLM, the server device 1 can appropriately evaluate the risk of regulatory violations due to emotional appeals that are not directly expressed.

[0058] In the verbalization process, the server device 1 may describe not only the physical descriptions of objects and people in images and videos, but also the impressions and emotions that viewers are likely to have. For example, in the verbalization process for advertising material that includes "before / after" images of a thin person, not only the physical description, such as "the image on the left shows a fat woman, and the image on the right shows a slim woman," but also the viewer's psychological reactions, such as "I admire the figure of the woman on the right" and "It gives me hope that I can dramatically change in a short period of time," are verbalized. This improves the accuracy of evaluating advertising expressions that affect human emotions and psychology compared to direct processing by multimodal AI.

[0059] In particular, the depiction of viewer emotions in the verbalization process of the server device 1 is effective in evaluating advertising materials that may evoke specific interpretations or emotions depending on the cultural background or social context, even if the image appears neutral on the surface. For example, for an advertisement in which a person in a particular posture or attire is displayed together with a product, by verbalizing the viewer's interpretation, such as "the impression is of a trustworthy expert" or "the impression is of authority and seems to guarantee the effectiveness of the product," it is possible to detect expressions that evoke the image of a doctor or expert, even if the person is not wearing a white coat. This makes it possible to identify advertising expressions that may give viewers a false impression, even if they do not directly violate regulations.

[0060] Furthermore, in the verbalization process, the server device 1 can describe a variety of impressions and emotions that viewers with different attributes (such as age, gender, and cultural background) may have. For example, the server device 1 verbalizes the differences in emotional responses of each viewer demographic, such as "young people are likely to perceive it as something they aspire to or idealize" and "middle-aged and older people are likely to feel that it arouses health concerns" for the same advertising material. This allows the server device 1 to detect advertising expressions that may have an inappropriate effect only on specific vulnerable groups, enabling more detailed advertisement evaluations tailored to the target demographic.

[0061] Furthermore, the server device 1's verbalization process can also describe emotional changes in the temporal context of an advertisement. For example, in an advertising video with a storyline that unfolds from problem presentation to resolution, the server device 1 verbalizes the flow of emotions, such as, "At first, the viewer feels anxious and worried, but after the product is introduced, the viewer feels a sense of relief and hope." This enables the server device 1 to evaluate regulatory violation risks that depend on the temporal context, such as indirect appeals to efficacy and effectiveness through storytelling, which cannot be detected by analyzing a single frame. This verbalized information on emotional changes is also used by the advertising rule determination unit 113 to select advertising rules, leading to the application of more appropriate evaluation criteria.

[0062] The server device 1 evaluates a combination of a plurality of media elements and detects a violation due to a combination such as displaying an image of a slim person while saying in voice, "Eating has these effects!"

[0063] In order to improve processing efficiency, the server device 1 also employs techniques in frame processing, such as thinning out images in which the same sentence is detected or similar images, by checking only one of the images, or conversely, combining multiple images to perform OCR.

[0064] The checks performed by the server device 1 include compliance with laws (such as the Pharmaceutical and Medical Device Act and the Specified Commercial Transactions Act), checking copyrights and registered trademarks, checking the materials used, whether or not kanji characters are used in emergency situations, compliance with platform terms and conditions, legality of the content, compliance with ethics, risk of online outrages, accuracy and reliability of the content (video fact check), video and audio quality, final checks of titles, thumbnails, and descriptions, privacy and personal information protection, advertising display rules and PR notation, overall review, timing checks (synchronization of audio and subtitles), and script checks.

[0065] Furthermore, the server device 1 incorporates several innovations to improve the accuracy and efficiency of advertising material evaluation. For example, by adopting a field-specific evaluation model in the evaluation process, the server device 1 can make highly accurate judgments regarding regulations and expressions specific to specific industries (e.g., pharmaceuticals, cosmetics, health foods, etc.). The server device 1 may also be equipped with a mechanism for accumulating past evaluation results and user feedback and continuously improving the evaluation model through machine learning. Additionally, to improve processing speed, the server device 1 may employ a two-stage evaluation method in which a pre-scan is performed at low resolution and only portions where problems are detected are closely examined at high resolution. The server device 1 may also be equipped with a function for pre-classifying the format and composition patterns of advertising materials and automatically selecting an evaluation flow optimized for each pattern, thereby enabling efficient evaluation of a wide variety of advertising materials.

[0066] In addition, the server device 1 may perform ensemble evaluation using multiple AI models to improve the reliability of the evaluation results, making it possible to detect with a high probability subtle expressions and regulatory violations that are difficult to detect with a single model. Furthermore, the server device 1 may incorporate metadata such as the advertiser's industry, advertising media, and target demographic into the evaluation process to understand the context of the material, thereby enabling the selection and evaluation of more appropriate advertising rules.

[0067] Another embodiment of the present invention is a real-time advertising video production support system. This system instantly detects problems during the production process of advertising videos and proposes corrections before the content is completed. The server device 1, which works in conjunction with digital tools used to create advertisements, such as video editing software, performs real-time evaluations at each stage of the editing process, including material placement, transition settings, and text generation, and has the ability to instantly highlight, alert, and suggest alternative expressions that may violate legal regulations. This reduces the risk of requiring major revisions after completion and improves production efficiency. For example, a warning is displayed and alternative expressions are suggested immediately after inserting audio material containing effects that violate the Pharmaceutical and Medical Device Act.

[0068] Another embodiment is an ad generation AI prompt optimization system. This system has a function to optimize the prompts themselves for the AI ​​that generates advertising materials. The server device 1 analyzes the original prompt entered by the user and automatically converts it into a legally compliant expression, while also proposing expressions that maintain or improve advertising effectiveness. For example, if the server device 1 analyzes the prompt "Create an advertisement for a supplement with excellent weight loss effects," it detects a possibility of violating the Pharmaceutical and Medical Device Act and automatically generates an alternative prompt such as "Create an advertisement that conveys the appeal of dietary supplements that support a healthy lifestyle." This increases the probability that the generated advertising materials will comply with legal regulations and significantly reduces the amount of work required for corrections.

[0069] Another possible embodiment is a system that supports multi-regional regulations. The server device 1 assumes advertising deployment in multiple countries and regions and has an evaluation function that takes into account differences in advertising regulations in each region. When analyzing advertising materials, the server device 1 evaluates them based on the advertising-related laws and industry guidelines of the respective countries and regions by specifying the country or region where the advertising will be deployed, and presents revision proposals for each region. For example, when evaluating advertising material information for health foods, the server device 1 provides evaluation results and revision proposals that take into account the rules that must be referenced in each region, such as the Pharmaceutical and Medical Device Act in Japan, FDA regulations in the United States, and country-specific and EU common regulations in the EU.

[0070] Further, an embodiment may include an advertisement target demographic suitability evaluation function. The server device 1 has a function for applying evaluation criteria according to the advertisement's target demographic (e.g., age, gender, interests). For example, particularly strict criteria may be applied to advertisements aimed at younger generations, and misleading expressions or excessive health claims may be more sensitively detected. In advertisements targeted at the elderly, expressions that exploit cognitive biases or elements that incite anxiety may be analyzed in detail. This not only ensures compliance with laws and regulations, but also supports the creation of advertisements that take social responsibility into consideration, making it possible to prevent misunderstandings and disadvantages for specific vulnerable demographics.

[0071] Another embodiment of the present invention is a predictive compliance assessment system. The server device 1 analyzes trends in changes to laws and regulations in each country based on historical information on past changes to laws and regulations, events that triggered changes, and similar legal amendments in other countries, and has the function of proactively detecting expressions that may become problematic in the future. The server device 1 evaluates expressions in areas that do not violate current regulations but are prone to stricter regulations (e.g., environmental claims, health claims), including future risks, based on past legal amendment patterns and examples of administrative guidance. For example, if a description of the effects of a specific ingredient is currently acceptable but may be subject to future regulation due to the accumulation of related scientific knowledge or growing social interest, the server device 1 warns of this and suggests more sustainable alternative expressions.

[0072] Further, an embodiment of context-dependent expression evaluation can also be mentioned. The server device 1 has a function of evaluating the overall impression of a combination of words or images, rather than a single word or image. For example, the server device 1 detects the possibility that the combination of the phrase "research suggests" and an "rising graph" is perceived as an exaggeration of scientific evidence. The server device 1 also analyzes the possibility that the use of specific music or colors, when combined with the displayed text, will be interpreted as suggesting a specific effect. By comprehensively evaluating the impression resulting from the interaction of multiple media elements such as text, images, audio, and video, the server device 1 can identify in advance advertising expressions that are misleading overall, even if they do not directly violate regulations, and propose corrections.

[0073] Another embodiment is a personalized advertisement evaluation. The server device 1 has a function of automatically generating multiple possible display patterns for personalized advertisements that display different content based on viewer profiles and evaluating regulatory compliance for each of them. The server device 1, for example, works in conjunction with an API of an advertising platform to extract or predict variations that may be displayed based on set personalization rules and verify compliance of each pattern with laws and regulations. This makes it possible to detect in advance regulatory-violating content that may be displayed only to some users and comprehensively manage regulatory risks.

[0074] Another embodiment is a system equipped with an emotion and impact prediction and analysis function. The server device 1 has a function to predict the psychological impact of an advertisement, such as its emotional impact and memorability, and to support the creation of advertisements that comply with legal regulations while maintaining a strong effect. For example, if an expression that elicits "surprise" violates regulations, the server device 1 proposes an alternative expression that has the same emotional impact but is legally acceptable based on neuroscience and psychological knowledge. This makes it possible to maximize advertising effectiveness while ensuring compliance, thereby achieving both an increase in the advertiser's brand value and a reduction in legal risk.

[0075] For example, the server device 1 may include a translation unit that translates text information. The translation unit receives a translation target language from a user. In this case, each of the processing units described above may perform each process using the translated text information acquired by the server device 1 before or after translation by the translation unit.

[0076] For example, the material information analysis unit 112 analyzes basic advertising information from translated text information obtained by the server device 1 before or after the translation unit translates the text information. The advertising rule determination unit 113 may determine advertising rules based on the basic advertising information and information about the country or region where the target language is used, or information about the country or region specified by the user. In this case, for example, if the text information obtained by the server device 1 is English and the target language obtained by the translation unit is Korean, and furthermore the basic advertising information analyzed by the material information analysis unit 112 indicates that the target product is food, the advertising rule determination unit 113 may determine, for example, advertising guidelines for food in Korea as the advertising rules.

[0077] Furthermore, the translation unit may change the words, sentences, and expressions themselves of the translation based on the basic advertising information analyzed by the material information analysis unit 112 or the advertising rules determined by the advertising rule determination unit 113. In this case, for example, if the target product analyzed by the material information analysis unit 112 is cosmetics for men in their twenties, the translation unit may change the translation to words, sentences, and expressions that are likely to appeal to men in their twenties. Similarly, for example, if the advertising rule determined by the advertising rule determination unit 113 is the Pharmaceutical and Medical Device Act, the translation unit may change the translation to one that does not include expressions that are prohibited from being used in pharmaceutical advertisements.

[0078] The translation unit may translate the information output by each processing unit into a language specified by the user. In this case, the presentation unit 116 presents the basic advertising information, advertising rules, evaluation results, correction material information, etc. translated by the translation unit to the user terminal 3.

[0079] Although the preferred embodiments of the present disclosure have been described in detail above with reference to the accompanying drawings, the technical scope of the present disclosure is not limited to such examples. It is clear that a person skilled in the art of the present disclosure can conceive of various modified or altered examples within the scope of the technical idea described in the claims, and it is understood that these also naturally fall within the technical scope of the present disclosure.

[0080] The devices described in this specification may be realized as a single device, or may be realized by multiple devices (e.g., cloud servers) some or all of which are connected via a network. For example, the CPU and storage device of the management server 20 may be realized by different servers connected to each other via a network.

[0081] The series of processes performed by the device described herein may be implemented using software, hardware, or a combination of software and hardware. A computer program for implementing each function of the management server 20 according to this embodiment may be created and installed on a PC or the like. A computer-readable recording medium on which such a computer program is stored may also be provided. Examples of the recording medium include a magnetic disk, an optical disk, a magneto-optical disk, and a flash memory. The computer program may also be distributed, for example, via a network, without using a recording medium.

[0082] Additionally, the processes described herein do not necessarily have to be performed in the order described, some process steps may be performed in parallel, additional process steps may be employed, and some process steps may be omitted.

[0083] Furthermore, the effects described herein are merely descriptive or exemplary and are not limiting. In other words, the technology according to the present disclosure may achieve other effects that will be apparent to those skilled in the art from the description of this specification, in addition to or in place of the above-described effects.

[0084] REFERENCE SIGNS LIST 1 Server device 2 Network 3 User terminal 101 CPU 102 Memory 103 Storage device 104 Communication interface 105 Input device 106 Output device 111 Material information acquisition unit 112 Material information analysis unit 113 Advertising rule determination unit 114 Evaluation unit 115 Edited material information generation unit 116 Presentation unit 131 Material information storage unit 132 Advertising rule information storage unit 133 Edited material information storage unit

Claims

1. An evaluation device that evaluates material information, which is material for advertising, comprising: an advertising rule determination unit that determines advertising rules related to advertisements in which the material information is used based on the material information; and an evaluation unit that evaluates the material information based on the advertising rules.

2. The evaluation device according to claim 1, wherein the material includes text information and image information, and the advertising rule determination unit analyzes the text information to determine the advertising rules.

3. The evaluation device according to claim 1 or 2, wherein the evaluation unit evaluates the material information based on the advertising rules.

4. The evaluation device according to claim 3, further comprising: a modified material information generation unit that generates non-problematic modified material information when the evaluation unit evaluates that the material information is problematic under the advertising rules.

5. The evaluation device according to claim 4, further comprising: a presentation unit that presents the problematic material information and the corrected material information to a user based on the evaluation result of the evaluation unit.

6. An evaluation program for evaluating advertising material information, the evaluation program causing a computer to execute an advertising rule determination step for determining advertising rules related to advertisements in which the material information is used based on the material, and an evaluation step for evaluating the material based on the advertising rules.

7. An evaluation method for evaluating advertising material information, the evaluation method comprising: a computer executing an advertising rule determination step for determining, based on the material, advertising rules related to an advertisement in which the material information is used; and an evaluation step for evaluating the material based on the advertising rules.

Images