Program, control method for information processing device, and information processing device

The mechanism identifies and manages AI-generated images to prevent their misleading posting on social networks by adding identification information and restricting or modifying them.

JP2025128644APending Publication Date: 2025-09-03CANON KK
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
JP2024025433
Authority / Receiving Office
JP · JP
Patent Type
Applications
Current Assignee / Owner
Filing Date
2024-02-22
Publication Date
2025-09-03

AI Technical Summary

Technical Problem

Existing image generation technologies using AI fail to identify and prevent the posting of fabricated images that may mislead viewers, allowing potentially inappropriate content to be shared on social networking sites.

Method used

A mechanism that determines if an image is AI-generated by adding identification information, and restricts or modifies the posting of such images to prevent misinformation.

Benefits of technology

Prevents the posting of AI-generated images that could mislead viewers by identifying and managing AI-generated content effectively.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 2025128644000001_ABST
    Figure 2025128644000001_ABST
Patent Text Reader

Abstract

To provide a mechanism for preventing the posting of an AI generated image that can be misrecognized by browsers.SOLUTION: In a communication system 100, a communication device 102 performs communication with an AI-generated image generation server 101 that generates an image using a generative AI. The AI image generation server 101 generates an AI-generated image according to a generation request received from the communication device 102; and transmits the AI-generated image to the communication device 102. The communication device 102 includes a program for an SNS application to use a SNS service. The communication device 102 performs, when it is determined that an image included in received post information is an AI-generated image, the processing of restricting the posting the post information.SELECTED DRAWING: Figure 1
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

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

[0002] In recent years, image generation technology using AI has become widespread, and individuals can now easily generate images using generative AI. This has made it possible for anyone to easily generate realistic images, but at the same time, there have been occasional cases where malicious users have used generative AI to generate images of fictitious incidents or accidents, which has become a problem.

[0003] Meanwhile, the use of social networking sites (SNSs), which allow users to post messages and images and interact with each other via the Internet, is increasing. With the increase in the use of such SNSs, there is a problem of inappropriate content being posted. To address this problem, a technique has been proposed for analyzing posted information and detecting information containing inappropriate content. For example, in Patent Document 1, features are extracted from character strings, images, and voices included in posted information, and the features are compared with information in a database to determine whether the posted information contains inappropriate content. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-39764 Summary of the Invention [Problem to be solved by the invention]

[0005] However, with the technology of Patent Document 1, even if the image included in the posted information is a fabricated image generated by a generation AI, if the image contains information that could actually occur, the image may not be determined to be inappropriate. In such cases, there is a concern that posted information containing AI-generated images fabricated from fictitious incidents or accidents may be made public, causing viewers to misinterpret the information.

[0006] The present invention aims to provide a mechanism that can prevent the posting of AI-generated images that may mislead viewers. [Means for solving the problem]

[0007] In order to achieve the above-mentioned object, the program of the present invention is an application program for using a service that publishes posted information, and is characterized in that it causes an information processing device equipped with the program to execute the steps of determining whether an image included in the received posted information is an image generated by a generation AI, and restricting the posting of the posted information when it is determined that the image included in the received posted information is an image generated by the generation AI. [Effects of the Invention]

[0008] According to the present invention, it is possible to prevent the posting of AI-generated images that may mislead viewers. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a block diagram schematically illustrating a configuration of a communication system including a communication device as an information processing device according to an embodiment of the present invention. [Figure 2] FIG. 2 is a block diagram showing the schematic configuration of the AI ​​image generation server of FIG. 1. [Figure 3] FIG. 2 is a block diagram illustrating a schematic configuration of the communication device of FIG. [Figure 4] FIG. 2 is a block diagram showing a schematic configuration of the SNS server of FIG. [Figure 5] 2 is a flowchart showing the steps of an image generation process performed by a generation AI in the communication system of FIG. 1. [Figure 6] 10 is a flowchart showing the procedure of a first posting control process executed by the communication system of FIG. [Figure 7] 4 is a diagram showing an example of a screen displayed on the user I / F unit of FIG. 3. FIG. [Figure 8] 10 is a flowchart showing the procedure of a second posting control process executed by the communication system of FIG. [Figure 9] FIG. 10 is a diagram for explaining a composite image generated by the process of S814 in FIG. 8. [Figure 10] 10 is a flowchart showing the procedure of a third posting control process executed by the communication system of FIG. [Figure 11] FIG. 11 is a diagram showing an example of posted information processed by the process in step S1014 of FIG. [Figure 12] 10 is a flowchart showing the procedure of a fourth posting control process executed by the communication system of FIG. [Figure 13] FIG. 13 is a diagram showing an example of posted information processed by the process in step S1214 of FIG. [Figure 14] 10 is a flowchart showing the procedure of a fifth posting control process executed by the communication system of FIG. [Figure 15] 10 is a flowchart showing the procedure of a posted information transmission process executed by the communication device of FIG. [Figure 16] 10 is a flowchart showing another procedure of the posted information transmission process executed by the communication device of FIG. DETAILED DESCRIPTION OF THE INVENTION

[0010] Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Note that the following embodiments do not limit the scope of the present invention as defined in the claims, and not all of the combinations of features described in the embodiments are necessarily essential to the solution of the present invention.

[0011] FIG. 1 is a block diagram showing a schematic configuration of a communication system 100 including a communication device 102 as an information processing device according to this embodiment.

[0012] 1, the communication system 100 is composed of an AI image generation server 101, a communication device 102, and an SNS server 103. The AI ​​image generation server 101, the communication device 102, and the SNS server 103 can communicate with each other via a network 104.

[0013] In the communication system 100, the communication device 102 makes a request to generate an image in accordance with instructions received from the user. The AI ​​image generation server 101 generates an image (hereinafter referred to as an "AI-generated image") using a generation AI in accordance with the generation request received from the communication device 102, and transmits the generated AI-generated image to the communication device 102. The user operates the communication device 102 to create posted information. This posted information includes, for example, an AI-generated image generated by the AI ​​image generation server 101. The communication device 102 transmits the posted information created by the user to the SNS server 103. The SNS server 103 is a server that provides an SNS service that makes posted information public. The SNS server 103 registers the received posted information in the posted information database 405 in Figure 4, which will be described later. This allows multiple users of the SNS service to view the posted information.

[0014] Figure 2 is a block diagram showing a schematic configuration of the AI ​​image generation server 101 in Figure 1. In Figure 2, the AI ​​image generation server 101 includes a CPU 201, a RAM 202, a ROM 203, a storage unit 204, a GPU 207, an AI-generated identification information addition unit 208, and a network I / F unit 209. These are connected to each other via a bus 210.

[0015] The CPU 201 controls the operation of the AI ​​image generation server 101 and operates based on a program stored in the RAM 202. The ROM 203 is a boot ROM that stores the system boot program and the like. The storage unit 204 is a non-volatile device such as an HDD or SSD that stores a trained model 205 and an AI image generation program 206 for implementing generative AI. Note that, although a configuration in which an arbitrary trained model 205 and an existing AI image generation program 206 such as Stable Diffusion are applied as the image generation means of the AI ​​image generation server 101 will be described, other image generation means may also be used. The trained model 205 and the AI ​​image generation program 206 are loaded into the RAM 202 and executed by the CPU 201. When the AI ​​image generation server 101 receives a generation request from the communication device 102 or the like, the GPU 207 executes an AI image generation process to generate an AI-generated image in accordance with instructions from the CPU 201.

[0016] The AI-generated identification information addition unit 208 adds AI-generated identification information to the AI-generated image generated in the AI ​​image generation process to identify it as an image generated by the generation AI. This AI-generated identification information includes information indicating that the image was generated by the generation AI, information about the trained model 205 and AI image generation program 206 used, information included in the generation request, etc. The AI-generated identification information may be in the form of metadata such as header information of the image file, or may be superimposed on the image itself as an invisible digital watermark or visible text. The AI-generated image with the AI-generated identification information added is transmitted to the communication device 102 via the network 104 or stored in the storage unit 204.

[0017] The network I / F unit 209 is connected to the network 104 and inputs and outputs various types of information via the network 104 .

[0018] Fig. 3 is a block diagram showing a schematic configuration of the communication device 102 in Fig. 1. The communication device 102 is a device capable of communicating with the AI ​​image generation server 101 and the SNS server 103, such as a smartphone, tablet terminal, or PC.

[0019] 3, the communication device 102 includes a CPU 301, a RAM 302, an SSD 303, a user I / F unit 304, and a network I / F unit 305. These are connected to one another via a bus 306.

[0020] The CPU 301 controls the operation of the communication device 102 and operates based on a program loaded in the RAM 302. The SSD 303 stores system programs and the like. The SSD 303 also stores an SNS application program. The SNS application is an application for using an SNS service provided by the SNS server 103. The user I / F unit 304 includes a display, touch panel, keyboard, mouse, etc., and performs input / output processing with the user. The network I / F unit 305 may be configured to perform wired communication or wireless communication. In this embodiment, the communication device 102 may also include a call unit for realizing a telephone function, a photographing unit for realizing a camera function, etc., in addition to the components shown in FIG. 3.

[0021] Fig. 4 is a block diagram showing a schematic configuration of the SNS server 103 in Fig. 1. In Fig. 4, the SNS server 103 includes a CPU 401, a RAM 402, a ROM 403, a storage unit 404, an AI-generated identification information analysis unit 406, and a network I / F unit 407. These are connected to each other via a bus 408.

[0022] The CPU 401 controls the operation of the SNS server 103 and operates based on a program stored in the RAM 402. The ROM 403 is a boot ROM and stores a system boot program and the like. The storage unit 404 is a storage unit that uses a non-volatile device, such as an SSD, HDD, or eMMC. The storage unit 404 stores a posted information database 405 for managing posted information received from devices such as the communication device 102. The AI-generated identification information analysis unit 406 analyzes an image to determine whether or not AI-generated identification information has been added by the AI-generated identification information addition unit 208. The network I / F unit 407 is connected to the network 104 and inputs and outputs various types of information via the network 104.

[0023] Next, the process of generating an image by the generation AI in the communication system 100 will be described.

[0024] Fig. 5 is a flowchart showing the procedure of image generation processing by the generation AI in the communication system 100 of Fig. 1. This image generation processing is performed by the communication device 102 and the AI ​​image generation server 101.

[0025] 5A is a flowchart showing the procedure of the generation request transmission process performed by the communication device 102. The generation request transmission process is realized by the CPU 301 of the communication device 102 executing a program stored in the SSD 303 or the like.

[0026] In FIG. 5(a), first, CPU 301 receives image generation parameters for generating an AI-generated image from the user (S501). The image generation parameters are used in known AI image generation techniques, such as keywords, sentences (prompts), and images related to the image to be generated. Next, CPU 301 generates a generation request for the AI-generated image based on the received image generation parameters and sends this generation request to AI image generation server 101 (S502). Then, this process ends.

[0027] 5(b) is a flowchart showing the steps of the image generation control process performed by the AI ​​image generation server 101. The image generation control process is realized by the CPU 201 of the AI ​​image generation server 101 executing a program stored in the ROM 203 or the memory unit 204.

[0028] In FIG. 5(b), when CPU 201 receives a generation request from communication device 102 via network 104 (S503), it controls GPU 207 to perform AI image generation processing (S504). This generates an AI-generated image corresponding to the image generation parameters input by the user. Next, CPU 201 controls AI-generated identification information addition unit 208 to add AI-generated identification information to the AI-generated image generated in S504, indicating that the image was generated by a generation AI (S505). Next, CPU 201 transmits the AI-generated image with the added AI-generated identification information to communication device 102 via network 104 (S506). Thereafter, this processing ends.

[0029] Next, a process for posting information in the communication system 100 will be described.

[0030] 6 is a flowchart showing the procedure of a first posting control process executed by the communication system 100 of FIG.

[0031] 6A is a flowchart showing the procedure of the posted information transmission process performed by the communication device 102. The posted information transmission process is realized by the CPU 301 of the communication device 102 executing a program of an SNS application stored in the SSD 303.

[0032] In FIG. 6(a), CPU 301 accepts posted information from a user (S601). The posted information includes at least any one of text data, audio data, image data, and video data. In this embodiment, CPU 301 accepts posted information on a posting screen 701 of FIG. 7(a) displayed on user I / F unit 304 of communication device 102. Posting screen 701 includes a character input section 702, a file selection section 703, and a post button 704. When a user selects a file in file selection section 703, information about the selected file is displayed on posting screen 701. As an example, in FIG. 7(a), image 705 selected by the user in file selection section 703 is displayed on posting screen 701. When a user presses post button 704, CPU 301 accepts posted information.

[0033] Next, CPU 301 transmits the received posted information to SNS server 103 (S602). SNS server 103 performs the posted information registration process shown in FIG. 6(b), which will be described later, and, for example, if the received posted information includes an AI-generated image, transmits a warning notification indicating this to communication device 102.

[0034] Next, the CPU 301 determines whether or not this warning notice has been received from the SNS server 103 (S603).

[0035] If it is determined in S603 that this warning notification has not been received from the SNS server 103, the posted information transmission process ends. If it is determined in S603 that this warning notification has been received from the SNS server 103, the CPU 301 performs a process to restrict the posting of the posted information accepted in S601. Specifically, the CPU 301 displays a warning notification screen 706 of FIG. 7(b) on the posting screen 701 displayed on the user I / F unit 304 (S604). The warning notification screen 706 is a screen that allows the user (poster) to select whether or not to post the posted information accepted in S601 as is, and includes a post button 707 and a cancel button 708.

[0036] Next, the CPU 301 notifies the SNS server 103 of information about the button pressed by the user, either the post button 707 or the cancel button 708 (S605), and the post information transmission process ends.

[0037] 6(b) is a flowchart showing the procedure of the posted information registration process performed by the SNS server 103. The posted information registration process is realized by the CPU 401 of the SNS server 103 executing a program stored in the ROM 403 or the storage unit 404.

[0038] 6(b), the CPU 401 receives posted information transmitted from the communication device 102 (S611). Next, the CPU 401 analyzes the received posted information and determines whether or not the posted information includes an image (S612).

[0039] If it is determined in S612 that the posted information contains an image, CPU 401 controls AI-generated identification information analysis unit 406 to determine whether the image contained in the posted information is an AI-generated image (S613). For example, if AI-generated identification information is attached to the image contained in the posted information, it is determined that the image contained in the posted information is an AI-generated image. On the other hand, if AI-generated identification information is not attached to the image contained in the posted information, it is determined that the image contained in the posted information is not an AI-generated image.

[0040] If it is determined in S613 that the image included in the posted information is an AI-generated image, CPU 401 transmits a warning notice indicating that the posted information received in S611 includes an AI-generated image to communication device 102 (S614). Communication device 102 that has received this warning notice displays warning notice screen 706 in S604 described above. Also, information about which of post button 707 and cancel button 708 was pressed by the user is notified to SNS server 103 in S605 described above.

[0041] When the CPU 401 receives information about the button pressed by the user from the communication device 102 (S615), it determines whether the button pressed by the user is the post button 707 or the cancel button 708 based on the received information (S616).

[0042] If it is determined in S616 that the button pressed by the user is cancel button 708, CPU 401 discards the posted information received in S611 (S617). As a result, the posted information received in S611 is not registered in posted information database 405, and this posted information is not made public. Thereafter, the posted information registration process ends.

[0043] If it is determined in S612 that the posted information does not include an image, the posted information registration process proceeds to S618. Also, if it is determined in S613 that the image included in the posted information is not an AI-generated image, the posted information registration process also proceeds to S618. Furthermore, if it is determined in S616 that the button pressed by the user is the post button 707, the posted information registration process also proceeds to S618.

[0044] In S618, the CPU 401 registers the posted information received in S611 in the posted information database. This makes the posted information public, and the poster and other users of the SNS service can view the registered posted information, for example, posted information 709 in FIG. 7(c). A message display section 710 of the posted information 709 displays the character string entered by the poster in the character input section 702, and a file display section 711 displays the image selected by the poster in the file selection section 703. The posted information 709 may also include an icon display section 712 that displays the poster's icon and a username display section 713 that displays the poster's name. When the processing of S618 is completed, the posted information registration processing ends.

[0045] According to the embodiment described above, if it is determined that the image included in the posted information received in S601 is an AI-generated image, a process for restricting the posting of this posted information is performed, thereby preventing the posting of AI-generated images that may mislead viewers.

[0046] Furthermore, in the above-described embodiment, a warning notification screen 706 is displayed on user I / F unit 304, prompting the poster to select whether or not to post the posted information received in S601. That is, if the posted information received in S601 includes an AI-generated image, the poster is given an opportunity to reconsider whether or not to post the posted information. This makes it possible to prevent the poster from posting an AI-generated image that unintentionally misleads viewers.

[0047] Furthermore, in the above-described embodiment, the determination of whether an image included in posted information is an AI-generated image is made based on whether AI-generated identification information is added to the image included in the received posted information. This makes it possible to easily determine whether an image included in posted information is an AI-generated image in a configuration in which AI image generation server 101 includes AI-generated identification information addition unit 208.

[0048] Although the present invention has been described using the above-mentioned embodiment, the present invention is not limited to the above-mentioned embodiment. For example, posted information may include multiple images. When posted information includes multiple images, notification may be provided as to which image is an AI-generated image.

[0049] Among AI-generated images, photographs are likely to give the impression of fictitious incidents or accidents, but illustrations are less likely to do so. Therefore, if the AI-generated image is an illustration in S613 described above, the posting of information containing such AI-generated images may not be restricted, and the process may proceed to S618.

[0050] Furthermore, in the above-described embodiment, the determination in S613 is based on the presence or absence of AI-generated identification information. However, this configuration is not limiting. For example, the determination in S613 may use an AI that receives an image included in the posted information as input and outputs a determination result indicating whether the image is an AI-generated image. Note that the SNS server 103 may have a trained model for implementing this AI. Alternatively, the SNS server 103 may transmit the image included in the posted information to an AI server that has a trained model for implementing this AI, and obtain the determination result from the AI ​​server. By using such an AI, it is possible to easily determine whether the image included in the posted information is an AI-generated image, even if the AI ​​image generation server 101 does not have the AI-generated identification information addition unit 208.

[0051] Furthermore, in this embodiment, the determination in S613 may be made based on predetermined information included in the AI-generated identification information. The predetermined information may be, for example, a probability or a score indicating the likelihood of the image being an AI-generated image. For example, in S613, CPU 401 converts the likelihood of the image being an AI-generated image into a percentage, and if the likelihood of a certain portion being an AI-generated image is equal to or greater than a threshold value N%, determines that the image included in the posted information is an AI-generated image. On the other hand, if the likelihood is less than the threshold value N%, determines that the image included in the posted information is not an AI-generated image. Note that threshold value N% may be stored in storage unit 404 in advance or may be included in the posted information. Using such a probability or score indicating the likelihood of the image being an AI-generated image, it is possible to determine whether or not the image included in the posted information is an AI-generated image.

[0052] Furthermore, in the above-described embodiment, the process of restricting posting of the posted information received in S601 involves displaying warning notification screen 706, which prompts the user to select whether or not to post the posted information as is. However, this is not the only possible configuration. For example, the process of restricting posting of the posted information received in S601 may involve displaying another warning notification screen on user I / F unit 304 to instruct the user to prohibit posting of the posted information. The other warning notification screen includes a message indicating that the posted information cannot be posted because it contains an AI-generated image, as well as cancel button 708, but does not include post button 707. This reliably prevents the posting of AI-generated images that may mislead viewers.

[0053] In the above-described embodiment, the posted information is restricted when the image included in the posted information is an AI-generated image, but the present invention is not limited to this configuration. For example, when the image included in the posted information is an AI-generated image, the posted information may be subjected to processing to prevent viewers from mistaking the image.

[0054] Fig. 8 is a flowchart showing the steps of a second posting control process executed by the communication system 100 of Fig. 1. The second posting control process is similar to the first posting control process described above, and the following will particularly describe the differences from the first posting control process described above. The second posting control process is also performed by the communication device 102 and the SNS server 103, similar to the first posting control process described above.

[0055] Fig. 8(a) is a flowchart showing the procedure of the posted information transmission process performed by the communication device 102. Like the posted information transmission process of Fig. 6(a) described above, the posted information transmission process of Fig. 8(a) is also realized by the CPU 301 of the communication device 102 executing the program of the SNS application stored in the SSD 303.

[0056] In FIG. 8A, steps S801 and S802, which are the same as steps S601 and S602 described above, are performed, and the posted information transmission process ends.

[0057] Fig. 8(b) is a flowchart showing the procedure of the posted information registration process performed by the SNS server 103. Like the posted information registration process of Fig. 6(b) described above, the posted information registration process of Fig. 8(b) is also realized by the CPU 401 of the SNS server 103 executing a program stored in the ROM 403 or the storage unit 404.

[0058] In FIG. 8(b), steps S811 to S813 are performed, which are the same processes as steps S611 to S613 described above.

[0059] If it is determined in S812 that the posted information does not contain an image, or if it is determined in S813 that the image contained in the posted information is not an AI-generated image, the posted information registration process of Fig. 8(b) proceeds to S815, which is the same process as S618 described above. Thereafter, the posted information registration process of Fig. 8(b) ends.

[0060] If it is determined in S813 that the image included in the posted information is an AI-generated image, the CPU 401 performs processing on the posted information to prevent viewers from misidentifying the image. Specifically, the CPU 401 adds additional information indicating that the image is an AI-generated image to the image determined to be an AI-generated image in the posted information (S814). The processing of S814 will now be described in detail. Image 901 in FIG. 9(a) is an example of an AI-generated image before the processing of S814. In S814, the CPU 401 superimposes additional information 902 indicating that the image 901 is an AI-generated image onto the image 901, as shown in FIG. 9(b). Note that, although FIG. 9(b) illustrates an example in which the character string "AI" is added as the additional information 902, other character strings or images may be added as additional information as long as the information allows the viewer to recognize that the image is an AI-generated image. Next, the posted information registration processing of FIG. 8(b) proceeds to S815, which is the same processing as S618 described above.

[0061] In the above-described embodiment, when an image included in received posted information is determined to be an AI-generated image, the posted information is processed to prevent viewers from mistaking it. In other words, when the received posted information includes an AI-generated image, the posted information is not posted in a state that could mislead viewers. This makes it possible to prevent the posting of AI-generated images that could mislead viewers.

[0062] Furthermore, in the above-described embodiment, in the processing, image 901 included in the received posted information is combined with additional information 902 indicating that image 901 is an AI-generated image. This makes it possible to make a viewer who views posted information including image 901 combined with additional information 902 aware that the posted information may contain information that is not factual, thereby preventing the viewer from misunderstanding something.

[0063] In this embodiment, as a processing step to prevent viewers from misunderstanding, additional information indicating that the image is an AI-generated image may be added to the property information of the image included in the received posted information. This makes it possible to make viewers aware that the posted information may contain information that is not factual, without editing the AI-generated image included in the posted information.

[0064] In addition, in this embodiment, as a processing step to prevent viewers from misunderstanding, a process may be performed in which a message indicating that the received posted information contains an AI-generated image is added to the posted information.

[0065] Fig. 10 is a flowchart showing the steps of a third posting control process executed by the communication system 100 of Fig. 1. The third posting control process is similar to the first and second posting control processes described above, and the following will particularly describe the differences from the first and second posting control processes described above. The third posting control process is also performed by the communication device 102 and the SNS server 103, similar to the first and second posting control processes described above.

[0066] Fig. 10(a) is a flowchart showing the procedure of the posted information transmission process performed by the communication device 102. Like the posted information transmission processes of Fig. 6(a) and Fig. 8(a) described above, the posted information transmission process of Fig. 10(a) is also realized by the CPU 301 of the communication device 102 executing a program of an SNS application stored in the SSD 303.

[0067] In FIG. 10A, steps S1001 and S1002, which are the same as steps S601 and S602 described above, are performed, and the posted information transmission process ends.

[0068] Fig. 10(b) is a flowchart showing the procedure of the posted information registration process performed by the SNS server 103. The posted information registration process of Fig. 10(b) is also realized by the CPU 401 of the SNS server 103 executing a program stored in the ROM 403 or the storage unit 404, similar to the posted information registration processes of Fig. 6(b) and Fig. 8(b) described above.

[0069] In FIG. 10(b), steps S1011 to S1013 are performed, which are the same processes as steps S611 to S613 described above.

[0070] If it is determined in S1012 that the posted information does not contain an image, or if it is determined in S1013 that the image contained in the posted information is not an AI-generated image, the posted information registration process of Fig. 10(b) proceeds to S1015, which is the same process as S618 described above. Thereafter, the posted information registration process of Fig. 10(b) ends.

[0071] If it is determined in S1013 that the image included in the posted information is an AI-generated image, CPU 401 processes the posted information to prevent viewers from mistaking it for an AI-generated image. Specifically, CPU 401 adds a message to the posted information indicating that the posted information contains an AI-generated image (S1014). Next, the posted information registration process of FIG. 10(b) proceeds to S1015, which is the same process as S618 described above. For example, when the posted information with the added message is registered in posted information database 405 in S1015, the poster and other users of the SNS service can view the posted information, for example, posted information 1101 in FIG. 11. In addition to character string 1102 entered by the poster, posted information 1101 includes message 1103 indicating that the posted information contains an AI-generated image. When the process of S1015 is completed, the posted information registration process of FIG. 10(b) ends.

[0072] In the above-described embodiment, a processing step for preventing viewers from mistaking the posted information is performed by adding a message to the posted information indicating that the posted information contains an AI-generated image. This allows viewers to immediately notice that the posted information may contain information that is not factual, without editing the AI-generated image included in the posted information.

[0073] In this embodiment, as a processing process to prevent viewers from misunderstanding, a process of excluding AI-generated images from posted information may be performed.

[0074] Fig. 12 is a flowchart showing the procedure of a fourth posting control process executed by the communication system 100 of Fig. 1. The fourth posting control process is similar to the first, second, and third posting control processes described above, and the following will particularly describe the differences from the first, second, and third posting control processes described above. The fourth posting control process is also performed by the communication device 102 and the SNS server 103, similar to the first, second, and third posting control processes described above.

[0075] Fig. 12(a) is a flowchart showing the procedure of the posted information transmission process performed by the communication device 102. Like the posted information transmission processes of Fig. 6(a), Fig. 8(a), and Fig. 10(a), the posted information transmission process of Fig. 12(a) is also realized by the CPU 301 of the communication device 102 executing a program of an SNS application stored in the SSD 303.

[0076] In FIG. 12A, steps S1201 and S1202, which are the same processes as steps S601 and S602 described above, are performed, and the posted information transmission process ends.

[0077] Fig. 12(b) is a flowchart showing the procedure of the posted information registration process performed by the SNS server 103. The posted information registration process of Fig. 12(b) is also realized by the CPU 401 of the SNS server 103 executing a program stored in the ROM 403 or the storage unit 404, similar to the posted information registration processes of Fig. 6(b), Fig. 8(b), and Fig. 10(b) described above.

[0078] In FIG. 12(b), steps S1211 to S1213 are performed, which are the same processes as steps S611 to S613 described above.

[0079] If it is determined in S1212 that the posted information does not contain an image, or if it is determined in S1213 that the image contained in the posted information is not an AI-generated image, the posted information registration process of Fig. 12(b) proceeds to S1215, which is the same process as S618 described above. Thereafter, the posted information registration process of Fig. 12(b) ends.

[0080] If it is determined in S1213 that the image included in the posted information is an AI-generated image, CPU 401 processes the posted information to prevent viewers from mistaking it for an AI-generated image. Specifically, CPU 401 removes the AI-generated image from the posted information (S1214). If the posted information includes multiple images, only the AI-generated image is removed from the multiple images. Next, the posted information registration process of FIG. 12(b) proceeds to S1215, which is the same process as S618 described above. For example, when the posted information from which the AI-generated image has been removed is registered in posted information database 405 in S1215, the poster and other users of the SNS service can view the posted information, for example, posted information 1301 in FIG. 13. In posted information 1301, the placement area of ​​the removed AI-generated image is displayed blank, as shown in 1302. When the process of S1015 is completed, the posted information registration process of FIG. 12(b) ends.

[0081] In the above-described embodiment, the processing to prevent viewers from misidentifying the posted information involves excluding AI-generated images from the posted information, thereby reliably preventing viewers from misidentifying the posted information due to AI-generated images.

[0082] In this embodiment, in S1214, the posted information may be subjected to a process of filling in the area of ​​the AI-generated image with the background color or another color.

[0083] In addition, in this embodiment, if an image included in posted information is an AI-generated image, it may be possible to control whether or not to process the posted information based on information about the source of the image.

[0084] For example, image super-resolution and color conversion using AI generation technology generate images based on the input image. Therefore, the generated images contain factual information, and even if information containing such images is posted as is, there is little chance that it will mislead third parties. On the other hand, among AI generation technologies, for example, Stable Diffusion generates images from strings of characters, and DCGAN generates images from noise. In this way, there is concern that AI-generated images generated by generative AI based on information other than images may be fictitious images that are different from the facts, and if information containing such images is posted as is, there is an extremely high possibility that it will mislead third parties.

[0085] In contrast, in this embodiment, when an image included in posted information is an AI-generated image, whether or not to apply processing to this posted information is controlled based on the generation source information of this image.

[0086] Fig. 14 is a flowchart showing the steps of a fifth posting control process executed by the communication system 100 of Fig. 1. The fifth posting control process is similar to the first, second, third, and fourth posting control processes described above, and the following will particularly describe the differences from the first, second, third, and fourth posting control processes described above. The fifth posting control process is also performed by the communication device 102 and the SNS server 103, similar to the first, second, third, and fourth posting control processes described above.

[0087] Fig. 14(a) is a flowchart showing the procedure of the posted information transmission process performed by the communication device 102. Like the posted information transmission processes of Fig. 6(a), Fig. 8(a), Fig. 10(a), and Fig. 12(a), the posted information transmission process of Fig. 14(a) is also realized by the CPU 301 of the communication device 102 executing a program of an SNS application stored in the SSD 303.

[0088] In FIG. 14A, steps S1401 and S1402, which are the same as steps S601 and S602 described above, are performed, and the posted information transmission process ends.

[0089] Fig. 14(b) is a flowchart showing the procedure of the posted information registration process performed by the SNS server 103. The posted information registration process of Fig. 14(b) is also realized by the CPU 401 of the SNS server 103 executing a program stored in the ROM 403 or the storage unit 404, similar to the posted information registration processes of Figs. 6(b), 8(b), 10(b), and 12(b) described above.

[0090] In FIG. 14(b), steps S1411 to S1413 are performed, which are the same processes as steps S611 to S613 described above.

[0091] If it is determined in S1412 that the posted information does not contain an image, or if it is determined in S1413 that the image contained in the posted information is not an AI-generated image, the posted information registration process of Fig. 14(b) proceeds to S1416, which is the same process as S618 described above. Thereafter, the posted information registration process of Fig. 14(b) ends.

[0092] If it is determined in S1413 that the image included in the posted information is an AI-generated image, the CPU 401 determines whether the image was generated from information other than an image based on the generation source information included in the AI-generated identification information of the image (S1414).

[0093] If it is determined in S1414 that the source of this image is not information other than an image, that is, if the source of this image is an image, the posted information registration process in FIG. 14(b) proceeds to S1416. Here, if the source of this image is an image, this image is an image generated by the generation AI based on an input image, that is, an image containing factual information. Therefore, even if posted information containing such an image is posted as is, it is unlikely to mislead third parties. Therefore, in S1416, the CPU 401 registers the posted information received in S1411 as is in the posted information database 405. Thereafter, the posted information registration process in FIG. 14(b) ends.

[0094] If it is determined in S1414 that the image was generated from information other than an image, the posted information registration process in FIG. 14(b) proceeds to S1415. Here, if the image was generated from information other than an image, the image is generated by the generation AI based on non-image information, such as text or noise. AI-generated images generated by the generation AI based on non-image information may be fabricated images depicting fictitious incidents or accidents that are not factual. If posted information containing such images is posted as is, there is a high possibility that it will mislead third parties. For this reason, in S1415, the CPU 401 performs a processing process on the posted information received in S1411 to prevent viewers from mistaking the image. The processing process is, for example, a process of combining additional information indicating that the image is an AI-generated image with the AI-generated image, as described above. The processing process is also a process of adding additional information indicating that the image is an AI-generated image as property information of the AI-generated image. The processing process is also a process of adding a message indicating that the AI-generated image is included. The processing process is also a process of excluding the AI-generated image from the posted information. Next, the process proceeds to S1516.

[0095] In the above-described embodiment, if it is determined that an image included in received posted information is an image generated by a generation AI using an input image as input, no processing is performed to prevent viewers from being misled. On the other hand, if it is determined that an image included in received posted information is an image generated by a generation AI using information other than the input image as input, processing is performed to prevent viewers from being misled. This makes it possible to perform the above-described processing on AI-generated images that may contain information that is not factual, while preventing excessive processing from being performed on AI-generated images that do not require processing.

[0096] In the above-described embodiment, the SNS server 103 determines whether an image included in posted information is an AI-generated image, but the present invention is not limited to this configuration. For example, the communication device 102 on which an SNS application is installed may determine whether the image is an AI-generated image.

[0097] Fig. 15 is a flowchart showing the steps of a posted information transmission process executed by the communication device 102 of Fig. 1. The posted information transmission process of Fig. 15 is similar to the first posted information control process described above, and the following will particularly describe the differences from the first posted information control process described above. Similar to the posted information transmission process of Fig. 6(a) described above, the posted information transmission process of Fig. 15 is also realized by the CPU 301 of the communication device 102 executing a program of an SNS application stored on the SSD 303.

[0098] 15, CPU 301 performs step S1501, which is the same process as step S601 described above. Next, CPU 301 analyzes the received posted information and determines whether or not the posted information includes an image (S1502).

[0099] If it is determined in S1502 that the posted information contains an image, CPU 301 determines whether the image included in the posted information is an AI-generated image (S1503). In the present embodiment, the determination in S1503 may be made based on the presence or absence of AI-generated identification information, as in S613 described above. Alternatively, the determination in S1503 may be made using an AI that receives an image included in the posted information as input and outputs a determination result indicating whether the image is an AI-generated image. In the present embodiment, communication device 102 may be equipped with a trained model for implementing this AI. Alternatively, communication device 102 may be configured to transmit the image included in the posted information to an AI server that has a trained model for implementing this AI, and obtain the determination result from the AI ​​server.

[0100] If it is determined in S1503 that the image included in the posted information is an AI-generated image, CPU 301 performs processing to restrict posting of the posted information accepted in S1501. Specifically, CPU 301 displays a warning notification screen 706 on posting screen 701 displayed on user I / F unit 304.

[0101] Next, CPU 301 accepts a button operation, such as the user pressing either post button 707 or cancel button 708 on warning notification screen 706 (S1505). Next, CPU 301 determines whether the button pressed by the user is post button 707 or cancel button 708 (S1506).

[0102] If it is determined in S1506 that the button pressed by the user is the cancel button 708, the CPU 301 discards the posted information accepted in S1501 (S1507), after which the process ends.

[0103] If it is determined in S1502 that the posted information does not contain an image, the process proceeds to S1508. Also, if it is determined in S1503 that the image included in the posted information is not an AI-generated image, the process proceeds to S1508. Furthermore, if it is determined in S1506 that the button pressed by the user is the post button 707, the process proceeds to S1508.

[0104] In S1508, CPU 301 registers the posted information received in S1501 in the posted information database, and then this process ends.

[0105] In the above-described embodiment, when communication device 102 with an SNS application installed determines that an image included in received posted information is an AI-generated image, processing is performed to restrict posting of the posted information, thereby preventing the posting of AI-generated images that may mislead viewers.

[0106] In addition, in the posted information transmission process of Figure 15 described above, as in the configuration described above, if the image determined to be an AI-generated image in S1503 is an illustration, the posting of posted information including this AI-generated image may not be restricted and the process may proceed to S1508.

[0107] Also, in the posted information transmission process of Figure 15 described above, similar to the configuration described above, the judgment in S1503 may be made based on predetermined information contained in the AI-generated identification information, for example, a probability or score indicating the likelihood of the image being AI-generated.

[0108] 15 described above, if an image included in the posted information is an AI-generated image, control may be performed to prohibit posting of the posted information, similar to the configuration described above. Specifically, in S1504, a message indicating that the posted information cannot be posted because it contains an AI-generated image, and another warning notification screen including a cancel button 708 but not a post button 707, are displayed on user I / F unit 304. When the user presses cancel button 708, the posted information accepted in S1501 is discarded (S1507). This reliably prevents the posting of an AI-generated image that may mislead viewers.

[0109] In addition, in this embodiment, if the image included in the posted information is an AI-generated image, the communication device 102 on which an SNS application is installed may process the posted information to prevent viewers from misunderstanding it.

[0110] Fig. 16 is a flowchart showing another procedure of the posted information transmission process executed by the communication device 102 of Fig. 1. Note that the posted information transmission process of Fig. 16 is similar to the posted information transmission process of Fig. 15 described above, and the following will particularly describe the differences from the posted information transmission process of Fig. 15 described above. Like the posted information transmission process of Fig. 15 described above, the posted information transmission process of Fig. 16 is also realized by the CPU 301 of the communication device 102 executing a program of an SNS application stored in SSD 303.

[0111] In FIG. 16, the CPU 301 performs steps S1601 to S1603, which are the same processes as steps S1501 to S1503 described above.

[0112] If it is determined in S1603 that the image included in the posted information is an AI-generated image, CPU 301 performs processing on the posted information received in S1601 to prevent viewers from mistaking it for an AI-generated image. Note that the processing is, for example, processing that combines additional information indicating that the image is an AI-generated image with the AI-generated image, as described above. Also, the processing is processing that adds additional information indicating that the image is an AI-generated image as property information of the AI-generated image. Also, the processing is processing that adds a message indicating that the AI-generated image is included. Also, the processing is processing that removes the AI-generated image from the posted information. Thereafter, CPU 301 registers the processed posted information in posted information database 405 (S1605), and this processing ends.

[0113] If it is determined in S1602 that the posted information does not contain an image, the process proceeds to S1605. Also, if it is determined in S1603 that the image contained in the posted information is not an AI-generated image, the process proceeds to S1605. In S1605, CPU 301 registers the posted information received in S1601 as is in posted information database 405. Then, the process ends.

[0114] In the above-described embodiment, when communication device 102, on which an SNS application is installed, determines that an image included in received posted information is an AI-generated image, the posted information is processed to prevent viewers from mistaking it. In other words, when received posted information includes an AI-generated image, the posted information is not posted in a state that could mislead viewers. This makes it possible to prevent the posting of AI-generated images that could mislead viewers.

[0115] 16 described above, if it is determined in S1603 that the image included in the posted information is an AI-generated image, CPU 301 may perform the same determination process as S1414 described above. If it is determined in this determination that the image was not generated from information other than an image, that is, if the image was generated from an image, the posted information transmission process in FIG. 16 proceeds to S1605. On the other hand, if it is determined in this determination that the image was generated from information other than an image, the posted information transmission process in FIG. 16 proceeds to S1604. By controlling in this manner, it is possible to perform the above-described processing on an AI-generated image that may contain information that is not factual, while preventing excessive processing from being performed on an AI-generated image that does not require processing.

[0116] In the present embodiment, when posted information includes an AI-generated image, the configuration has been described in which the posting of the posted information is restricted, the posting of the posted information is prohibited, or the posted information is processed. However, the present invention is not limited to this configuration. For example, when posted information includes voice data generated by a generation AI, the posting of the posted information may be restricted, the posting of the posted information may be prohibited, or the posted information may be processed. This processing may be, for example, a process of adding information indicating that the posted information was generated by a generation AI as property information of the voice data, or a process of adding a message to the posted information indicating that the posted information includes voice data generated by a generation AI.

[0117] The present invention can also be realized by supplying a program that realizes one or more functions of the above-described embodiments to a system or device via a network or a storage medium, and having one or more processors in the computer of the system or device read and execute the program.The present invention can also be realized by a circuit (e.g., ASIC) that realizes one or more functions.

[0118] The disclosure of this embodiment includes the following configurations and methods. (Configuration 1) A program for an application for using a service that publishes posted information, characterized in that the program causes an information processing device equipped with the program to execute the steps of: determining whether an image included in the received posted information is an image generated by a generation AI; and restricting the posting of the posted information when it is determined that the image included in the received posted information is an image generated by the generation AI. (Configuration 2) The program described in Configuration 1, characterized in that in the restricting step, a screen is displayed on the display means of the information processing device, allowing the poster of the posted information to choose whether or not to post the received posted information. (Configuration 3) The program described in Configuration 1, characterized in that in the restricting step, a screen is displayed on the display means of the information processing device to instruct the poster of the posted information to stop posting the accepted posted information. (Configuration 4) A program for an application for using a service that publishes posted information, the program causing an information processing device equipped with the program to execute the steps of: determining whether an image included in the received posted information is an image generated by a generation AI; and, if it is determined that the image included in the received posted information is an image generated by the generation AI, performing processing on the posted information to prevent viewers from misunderstanding it. (Configuration 5) The program described in Configuration 4, characterized in that in the processing, additional information indicating that the image contained in the received posting information is generated by a generation AI is synthesized with the image contained in the received posting information. (Configuration 6) The program described in Configuration 4, characterized in that in the processing, additional information indicating that the image contained in the received posting information is generated by a generation AI is added as property information of the image. (Configuration 7) The program described in Configuration 4, characterized in that in the processing process, a message indicating that the received posted information contains an image generated by the generation AI is added to the received posted information. (Configuration 8) The program according to Configuration 4, wherein the processing process excludes images generated by the generation AI from the received posted information. (Configuration 9) A program described in any one of configurations 4 to 8, characterized in that an information processing device equipped with the program does not execute the process of performing the processing when it is determined that an image included in the accepted posted information is an image generated by the generation AI using an input image as input, and executes the process of performing the processing when it is determined that an image included in the accepted posted information is an image generated by the generation AI using information other than the input image as input. (Configuration 10) A program described in any one of configurations 1 to 9, characterized in that the image generated by the generation AI is attached with identification information indicating that it was generated by the generation AI, and the determining process is performed based on whether the identification information is attached to the image included in the received posting information. (Configuration 11) The program described in any one of configurations 1 to 9, characterized in that the determination process is performed using an AI that takes an image included in the received posted information as input and outputs a determination result indicating whether or not the image was generated by the generation AI. [Explanation of symbols]

[0119] 102 Communication equipment 301 CPU 304 User I / F section 706 Warning notification screen 901 images 902 Additional Information 1103 Message

Claims

1. An application program for using a service that publishes posted information, An information processing device having the program, determining whether an image included in the received posted information is an image generated by a generation AI; and restricting the posting of the posted information when it is determined that the image contained in the received posted information is an image generated by the generation AI.

2. 2. The program according to claim 1, wherein in the restricting step, a screen is displayed on the display means of the information processing device, which allows the poster of the posted information to select whether or not to post the received posted information.

3. 2. The program according to claim 1, wherein the restricting step displays a screen on a display means of the information processing device that prompts a poster of the received posted information to stop posting the posted information.

4. An application program for using a service that publishes posted information, An information processing device having the program, determining whether an image included in the received posted information is an image generated by a generation AI; A program characterized by executing a process of processing the posted information to prevent viewers from misunderstanding it if it is determined that the image contained in the accepted posted information is an image generated by the generation AI.

5. The program according to claim 4, characterized in that in the processing, additional information indicating that the image contained in the received posted information is generated by a generation AI is synthesized with the image.

6. The program described in claim 4, characterized in that in the processing process, additional information indicating that the image contained in the accepted posting information is generated by a generation AI is added as property information of the image.

7. The program according to claim 4, characterized in that, in the processing process, a message indicating that the received posted information contains an image generated by the generation AI is added to the received posted information.

8. The program according to claim 4, wherein the processing step includes excluding images generated by the generation AI from the received posted information.

9. An information processing device having the program, When it is determined that the image included in the received posting information is an image generated by the generation AI using an input image as input, the step of performing the processing is not executed, The program described in claim 4, characterized in that when it is determined that the image contained in the received posted information is an image generated by the generation AI using information other than the input image as input, the program executes the process of performing the processing.

10. Identification information indicating that the image was generated by the generation AI is added to the image, 5. The program according to claim 1, wherein the determining step is performed based on whether or not the identification information is added to an image included in the received posted information.

11. The program described in claim 1 or 4, characterized in that the determination process is performed using AI that inputs an image contained in the accepted posted information and outputs a determination result indicating whether or not the image was generated by the generation AI.

12. A control method for an information processing device having an application program for using a service that publishes posted information, determining whether an image included in the received posted information is an image generated by a generation AI; A control method for an information processing device, characterized by comprising a step of restricting the posting of the posted information when it is determined that an image contained in the received posted information is an image generated by the generation AI.

13. A control method for an information processing device having an application program for using a service that publishes posted information, determining whether an image included in the received posted information is an image generated by a generation AI; A control method for an information processing device characterized by comprising a step of, when it is determined that an image contained in the received posted information is an image generated by the generation AI, processing the posted information to prevent viewers from misunderstanding it.

14. An information processing device having an application program for using a service that publishes posted information, A means for determining whether an image included in the received posted information is an image generated by a generation AI; An information processing device characterized by having a means for restricting the posting of the posted information when it is determined that an image contained in the received posted information is an image generated by the generation AI.

15. An information processing device having an application program for using a service that publishes posted information, A means for determining whether an image included in the received posted information is an image generated by a generation AI; An information processing device characterized by having a means for processing the posted information to prevent viewers from misunderstanding it when it is determined that an image contained in the accepted posted information is an image generated by the generation AI.

Citation Information

Patent Citations

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

    JP2022039764A