Resetting method and device for AI generation content, storage medium and program product

Abstract

The invention provides an AI generation content resetting method and device, a storage medium and a program product, and the method comprises the steps: obtaining to-be-processed artificial intelligence AI generation content in response to a resetting instruction; extracting generation process information of the AI generation content; generating watermark information corresponding to the AI generation content according to the generation process information; and embedding the watermark information into the AI generation content to obtain a reset target AI generation content. According to the method and the device, the AI generation content is reset under the condition that the visual or auditory effect of the AI generation content is not changed, and the traceability and the safety of the AI generation content are improved.

Description

Technical Field

[0001] The present application relates to the field of information processing technology, and in particular to a method, device, storage medium and program product for resetting AI-generated content. Background Art

[0002] Artificial Intelligence (AI) is the study of the technology of using computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and tasks that usually require human intelligence are performed by computer systems or machines. These tasks include but are not limited to learning, reasoning, problem solving, perception, and language understanding. Artificial intelligence technology is widely used in various fields, such as natural language processing, computer vision, robotics, and autonomous driving.

[0003] AI-generated content refers to content in various forms such as text, images, audio, and video that are automatically generated using artificial intelligence algorithms and models. With the rapid development and widespread application of AI generation technology, different types of AI generation models and software tools are emerging in an endless stream. At present, different types of AI generation software models and software tools are prone to information security issues such as data leakage and unclear ownership in the generation, transmission, and storage application stages. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to provide a method, device, storage medium and program product for resetting AI-generated content, which can reset AI-generated content without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content.

[0005] In a first aspect, an embodiment of the present application provides a method for resetting AI-generated content, comprising: in response to a reset instruction, obtaining artificial intelligence AI-generated content to be processed; extracting generation process information of the AI-generated content; generating watermark information corresponding to the AI-generated content according to the generation process information; embedding the watermark information into the AI-generated content to obtain the reset target AI-generated content.

[0006] In one embodiment, before obtaining the artificial intelligence AI-generated content to be processed in response to the reset instruction, it also includes: determining the source information of the AI-generated content; judging whether the AI-generated content comes from a trusted system based on the source information; if the AI-generated content does not come from a trusted system, triggering the reset instruction regarding the AI-generated content.

[0007] In one embodiment, the extracting the generation process information of the AI-generated content includes: inputting the AI-generated content into a preset recognition model so that the preset recognition model outputs the generation process information of the AI-generated content; wherein the preset recognition model is obtained by training a neural network model based on a first sample set, and the first sample set includes first sample content generated based on AI, and the first sample content is marked with corresponding generation process information.

[0008] In one embodiment, extracting the generation process information of the AI-generated content includes: parsing metadata of the AI-generated content, and determining the generation process information according to the metadata.

[0009] In one embodiment, generating the watermark information corresponding to the AI ​​generated content according to the generation process information includes: converting the information format of the generation process information into an information format specified by the target system; and generating the watermark information corresponding to the AI ​​generated content according to the generation process information after format conversion.

[0010] In one embodiment, generating watermark information corresponding to the AI ​​generated content according to the generation process information includes: extracting a feature code of the generation process information, wherein the feature code is used to characterize the uniqueness of the AI ​​generated content; and generating watermark information corresponding to the AI ​​generated content according to the feature code.

[0011] In one embodiment, the embedding of the watermark information into the AI ​​generated content to obtain the reset target AI generated content includes: determining the embedding position of the watermark information in the AI ​​generated content; determining the embedding strength of the watermark information according to the AI ​​generated content, wherein the embedding strength is less than or equal to a preset strength threshold; and embedding the watermark information into the AI ​​generated content according to the embedding strength to obtain the reset target AI generated content.

[0012] In one embodiment, before generating the watermark information corresponding to the AI ​​generated content according to the generation process information, it also includes: verifying the data integrity of the AI ​​generated content; and / or verifying whether the AI ​​generated content has been tampered with; when it is determined that the AI ​​generated content is complete and that the AI ​​generated content has not been tampered with, executing the step of generating the watermark information corresponding to the AI ​​generated content according to the generation process information.

[0013] In one embodiment, the verification of the data integrity of the AI-generated content includes: determining whether each field in the preset field set has matching information in the generation process information; if each field in the preset field set has matching information in the generation process information, determining that the AI-generated content is complete; if there are fields in the preset field set that have no matching information in the generation process information, determining that the AI-generated content is incomplete.

[0014] In one embodiment, the verifying whether the AI ​​generated content has been tampered with includes: using a preset detection model to identify whether the AI ​​generated content has been tampered with, wherein the preset detection model is obtained by training a neural network model based on a second sample set, and the second sample set includes a second sample content pair generated based on AI, and the second sample content pair includes original sample content that has not been tampered with and sample content that has been tampered with.

[0015] In a second aspect, an embodiment of the present application provides a method for resetting AI-generated content about a product, comprising: in response to a reset instruction, obtaining artificial intelligence (AI)-generated content to be processed, the AI-generated content including product information; extracting generation process information of the AI-generated content; generating watermark information corresponding to the AI-generated content according to the generation process information; embedding the watermark information into the AI-generated content to obtain the reset target AI-generated content.

[0016] In a third aspect, an embodiment of the present application provides a device for resetting AI-generated content, including:

[0017] An acquisition module, used for acquiring artificial intelligence AI generated content to be processed in response to a reset instruction;

[0018] An extraction module, used to extract generation process information of the AI-generated content;

[0019] A generation module, used to generate watermark information corresponding to the AI ​​generated content according to the generation process information;

[0020] The embedding module is used to embed the watermark information into the AI ​​generated content to obtain the reset target AI generated content.

[0021] In one embodiment, it also includes: a judgment module, which is used to determine the source information of the AI-generated content before obtaining the artificial intelligence AI-generated content to be processed in response to the reset instruction; judge whether the AI-generated content comes from a trusted system according to the source information; if the AI-generated content does not come from a trusted system, trigger the reset instruction regarding the AI-generated content.

[0022] In one embodiment, an extraction module is used to input the AI-generated content into a preset recognition model so that the preset recognition model outputs the generation process information of the AI-generated content; wherein the preset recognition model is obtained by training a neural network model based on a first sample set, and the first sample set includes a first sample content generated based on AI, and the first sample content is marked with the corresponding generation process information.

[0023] In one embodiment, the extraction module is used to parse the metadata of the AI-generated content and determine the generation process information based on the metadata.

[0024] In one embodiment, the generation module is used to convert the information format of the generation process information into an information format specified by the target system; and generate watermark information corresponding to the AI ​​generated content according to the generation process information after format conversion.

[0025] In one embodiment, the generation module is used to extract the feature code of the generation process information, where the feature code is used to characterize the uniqueness of the AI-generated content; and generate watermark information corresponding to the AI-generated content according to the feature code.

[0026] In one embodiment, an embedding module is used to determine the embedding position of the watermark information in the AI ​​generated content; determine the embedding strength of the watermark information according to the AI ​​generated content, and the embedding strength is less than or equal to a preset strength threshold; embed the watermark information into the AI ​​generated content according to the embedding strength to obtain the reset target AI generated content.

[0027] In one embodiment, it also includes: a verification module, which is used to verify the data integrity of the AI ​​generated content before generating the watermark information corresponding to the AI ​​generated content according to the generation process information; and / or, to verify whether the AI ​​generated content has been tampered with; when it is determined that the AI ​​generated content is complete and that the AI ​​generated content has not been tampered with, the step of generating the watermark information corresponding to the AI ​​generated content according to the generation process information is executed.

[0028] In one embodiment, the verification module is specifically used to determine whether each field in the preset field set has matching information in the generation process information; if each field in the preset field set has matching information in the generation process information, it is determined that the AI ​​generated content is complete; if there are fields in the preset field set that have no matching information in the generation process information, it is determined that the AI ​​generated content is incomplete.

[0029] In one embodiment, the verification module is specifically used to use a preset detection model to identify whether the AI-generated content has been tampered with, wherein the preset detection model is obtained by training a neural network model based on a second sample set, and the second sample set includes a second sample content pair generated based on AI, and the second sample content pair includes the original sample content that has not been tampered with and the sample content that has been tampered with.

[0030] In a fourth aspect, an embodiment of the present application provides an electronic device, including:

[0031] at least one processor; and

[0032] a memory communicatively coupled to the at least one processor;

[0033] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the electronic device to execute the method described in any one of the above aspects.

[0034] In a fifth aspect, an embodiment of the present application provides a cloud device, including:

[0035] at least one processor; and

[0036] a memory communicatively coupled to the at least one processor;

[0037] The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the cloud device to execute the method described in any one of the above aspects.

[0038] In a sixth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer-executable instructions. When a processor executes the computer-executable instructions, the method described in any one of the above aspects is implemented.

[0039] In a seventh aspect, an embodiment of the present application provides a computer program product, including a computer program, which implements the method described in any of the above aspects when executed by a processor.

[0040] The AI-generated content resetting method, device, storage medium and program product provided in the embodiments of the present application extract the generation process information of the AI-generated content and generate corresponding watermark information based on the generation process information. Since the generation process information can characterize the uniqueness of the AI-generated content, the corresponding watermark information can uniquely identify the AI-generated content. After the watermark information is embedded in the AI-generated content, the reset target AI-generated content can have an accurate unique identification, thereby resetting the AI-generated content without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application. Obviously, the drawings described below are some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0042] Figure 1 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application;

[0043] Figure 2 A schematic diagram of an application scenario of an AI-generated content reset system provided in an embodiment of the present application;

[0044] Figure 3 A flowchart of a method for resetting AI-generated content provided in an embodiment of the present application;

[0045] Figure 4 A flowchart of a method for resetting AI-generated content provided in an embodiment of the present application;

[0046] Figure 5 A flowchart of a method for resetting AI-generated content for a product provided in an embodiment of the present application;

[0047] Figure 6 A schematic diagram of the structure of a device for resetting AI-generated content provided in an embodiment of the present application;

[0048] Figure 7 A schematic diagram of the structure of a cloud device provided in an embodiment of the present application.

[0049] The above drawings have shown clear embodiments of the present application, which will be described in more detail later. These drawings and text descriptions are not intended to limit the scope of the present application in any way, but to illustrate the concept of the present application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0050] Here, exemplary embodiments will be described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.

[0051] The term "and / or" in this article is used to describe the association relationship of associated objects, specifically indicating that there may be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone.

[0052] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0053] In order to clearly describe the technical solution of the embodiment of the present application, the terms involved in the present application are first defined:

[0054] AI: Artificial Intelligence.

[0055] AI-generated content: AI-generated content refers to data content generated by artificial intelligence algorithms, including but not limited to text, images, audio, and video.

[0056] Digital watermarking: refers to embedding specific information into a digital signal, which may be audio, image, or video. If a signal with a digital watermark is copied, the embedded information will also be copied.

[0057] Invisible Watermarking: Invisible Watermarking is information embedded in digital media (such as images, videos, audio) to protect copyright, verify authenticity, or convey other secret information. The invisible watermark is invisible or imperceptible during normal viewing or use, but can be extracted and verified through specific algorithms or tools.

[0058] GPU: graphics processing unit, graphics processor.

[0059] EXIF: Exchangeable Image File Format, exchangeable image file format.

[0060] The method for resetting AI-generated content in the embodiment of the present application can be applied to any field involving AI data processing.

[0061] Artificial Intelligence (AI) is the study of the technology of using computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, planning, etc.), and tasks that usually require human intelligence are performed by computer systems or machines. These tasks include but are not limited to learning, reasoning, problem solving, perception, and language understanding. Artificial intelligence technology is widely used in various fields, such as natural language processing, computer vision, robotics, and autonomous driving.

[0062] AI-generated content refers to content in various forms such as text, images, audio, and video that are automatically generated using artificial intelligence algorithms and models. With the rapid development and widespread application of AI generation technology, different types of AI generation models and software tools are emerging in an endless stream. At present, different types of AI generation software models and software tools are prone to information security issues such as data leakage and unclear ownership in the generation, transmission, and storage application stages.

[0063] Taking the e-commerce platform scenario as an example, merchants on the e-commerce platform can use AI big models to generate different product images for product updates. However, the AI ​​product images uploaded by merchant users may come from different AI systems. The generation, transmission, and storage of AI product images are prone to data leakage and unclear ownership, resulting in great information security risks for AI-generated content.

[0064] In order to solve at least one of the above problems, an embodiment of the present application provides a reset solution for AI-generated content, by extracting the generation process information of the AI-generated content and generating corresponding watermark information based on the generation process information. Since the generation process information can characterize the uniqueness of the AI-generated content, the corresponding watermark information can uniquely identify the AI-generated content. After the watermark information is embedded in the AI-generated content, the reset target AI-generated content can have an accurate unique identification, thereby resetting the AI-generated content without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content.

[0065] Some embodiments of the present application are described in detail below in conjunction with the accompanying drawings. In the case where there is no conflict between the embodiments, the following embodiments and the features in the embodiments can be combined with each other. In addition, the step sequence in the following method embodiments is only an example and not a strict limitation.

[0066] like Figure 1 As shown, this embodiment provides an electronic device 1, including: at least one processor 11 and a memory 12, Figure 1A processor is taken as an example. The processor 11 and the memory 12 are connected via a bus 10. The memory 12 stores instructions that can be executed by the processor 11, and the instructions are executed by the processor 11 so that the electronic device 1 can execute all or part of the process of the method in the following embodiment to achieve resetting the AI-generated content without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content.

[0067] In one embodiment, the electronic device 1 may be a mobile phone, a tablet computer, a laptop computer, a desktop computer, or a large computing system composed of multiple computers.

[0068] Figure 2 A schematic diagram of an application scenario 200 of a system for resetting AI-generated content provided in an embodiment of the present application.

[0069] like Figure 2 As shown, the system includes: a server 210 and a terminal 220, wherein:

[0070] Server 210 may be a data platform that provides AI-generated content resetting services, such as an e-commerce platform. In actual scenarios, an e-commerce platform may have multiple servers 210. Figure 2 In the figure, one server 210 is taken as an example.

[0071] The terminal 220 may be a computer, mobile phone, tablet or other device used by the user to log in to the e-commerce platform. There may also be multiple terminals 220. Figure 2 Two terminals 220 are taken as an example for illustration.

[0072] The terminal 220 and the server 210 can transmit information via the Internet, so that the terminal 220 can access the data on the server 210. The terminal 220 and / or the server 210 can be implemented by the electronic device 1.

[0073] The reset solution of AI generated content in the embodiment of the present application can be deployed on the server 210, or on the terminal 220, or partially on the server 210 and partially on the terminal 220. In actual scenarios, it can be selected based on actual needs, and this embodiment does not limit it.

[0074] When the reset solution of the AI ​​generated content is fully or partially deployed on the server 210, a calling interface can be opened to the terminal 220 to provide algorithm support to the terminal 220.

[0075] The method provided in the embodiment of the present application can be implemented by executing the corresponding software code by the electronic device 1, and can be implemented by exchanging data with the server. The electronic device 1 can be a local terminal device. When the method is run on the server, the method can be implemented and executed based on the cloud interaction system, wherein the cloud interaction system includes a server and a client device.

[0076] In a possible implementation, the method provided in the embodiment of the present application provides a graphical user interface through a terminal device, wherein the terminal device can be the local terminal device mentioned above, or can be the client device in the cloud interaction system mentioned above.

[0077] Please see Figure 3 , which is a method for resetting AI-generated content in one embodiment of the present application, which can be Figure 1 The electronic device 1 shown is used to perform and can be applied to Figure 2 In the reset application scenario of AI-generated content shown in , the AI-generated content can be reset without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content. In this embodiment, the terminal 220 is used as an execution end as an example, and the method includes the following steps:

[0078] Step 301: In response to a reset instruction, obtain artificial intelligence AI generated content to be processed.

[0079] In this step, the reset instruction refers to the reset instruction of the AI ​​generated content to be processed. The reset instruction can be actively triggered by the user, such as after the merchant user of the e-commerce platform uploads the AI ​​product image, the reset instruction of the AI ​​product image is actively triggered to make the AI ​​product image recognizable after being reset. The reset instruction can also be automatically triggered by the system. For example, after the e-commerce platform detects the AI ​​product image created or uploaded by the user online, if the AI ​​product image has not been reset by the information of this system, or is an untrustworthy AI generated content, the reset instruction for the AI ​​product image can be automatically triggered.

[0080] AI-generated content refers to content generated based on artificial intelligence algorithms, including but not limited to text, images, audio, and video. The AI-generated content to be processed can be AI-generated content specified by the user, such as AI-generated product images or text uploaded by the user on the e-commerce platform. The AI-generated content can come from different systems.

[0081] The AI-generated content to be processed can also be the AI-generated content produced by users online. The system can automatically obtain the AI-generated content to be processed, reduce the need for manual intervention, and improve the efficiency and accuracy of data content processing. Users produce AI-generated content online means that the system uses artificial intelligence algorithms and cloud computing technology to process user input in real time and quickly generate digital content that matches the user input. The system can provide AI generation functions so that users can produce AI works online, and can provide highly personalized content generation services based on users' historical data and preferences. For example, e-commerce platforms can be integrated with artificial intelligence algorithm modules for users to produce AI-generated content online, and the system can obtain AI-generated content produced online in real time. By analyzing user behavior and needs, the system can generate content that better meets user expectations, thereby improving user satisfaction.

[0082] The AI-generated content here can be user-oriented images, audio, video, text and other content generated by AI, or it can be the underlying information of these images, audio, video, text and other content, such as the corresponding source code information, content address, data feature information of the content, etc.

[0083] In one embodiment, before step 301, the method further includes: determining source information of the AI-generated content. Determining whether the AI-generated content comes from a trusted system based on the source information. If the AI-generated content does not come from a trusted system, triggering a reset instruction for the AI-generated content.

[0084] In this embodiment, the source information of AI-generated content includes but is not limited to the source system of AI-generated content, the original technical language used, the creation platform and other information. A trusted system refers to an AI-generated system that has been security-certified, and the data storage format of the trusted system is known and recognizable. Before resetting the data, you can first determine whether the source of the AI-generated content is safe. If the AI-generated content comes from an untrusted system, it means that the AI-generated content has information security risks, and information leakage may occur during transmission and use. Here, a reset instruction for the AI-generated content can be automatically triggered to enable the system to reset its data. The reset AI-generated content can reset the AI-generated content without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and information security of the AI-generated content.

[0085] Optionally, the judgment method of the trusted system can adopt a system whitelist method, in which the information of the secure AI-generated system is pre-configured, and the source information of the AI-generated content to be processed is compared with the system information in the whitelist. If the source information of the AI-generated content to be processed exists in the whitelist, it is determined that the AI-generated content to be processed comes from a trusted system; otherwise, it is determined that the AI-generated content to be processed does not come from a trusted system, and a reset instruction for the AI-generated content to be processed is automatically triggered.

[0086] Optionally, the judgment method of the trusted system can also adopt an online search and matching method. For example, the source information of the AI-generated content to be processed is used as the search keyword, and a matching AI-generated system is searched online. Then, an information security verification is performed on the AI-generated system. If the verification passes, it is determined that the AI-generated content to be processed is from a trusted system. Otherwise, it is determined that the AI-generated content to be processed is not from a trusted system, and a reset instruction for the AI-generated content to be processed is automatically triggered.

[0087] Optionally, a machine learning model can be built to extract features of AI-generated content and collect relevant information about content generation, such as the technology used in AI-generated content, generation time, confirmation time, written content (combination of business agreements and / or national standards), technical processes, etc. Then, based on the features of AI-generated content, it is determined whether the AI-generated content comes from an untrusted system.

[0088] The rule engine can be configured to compare with the blacklist database, which stores information about untrusted systems, and determine the source of AI-generated content through machine learning models and rule engines. If AI content generated by an untrusted system is identified, it can be automatically marked to trigger further processing.

[0089] Qualification information can be directly filed for AI-generated content from trusted systems.

[0090] Step 302: Extracting the generation process information of the AI ​​generated content.

[0091] In this step, the generation process information refers to the relevant information of the AI-generated content to be processed during the generation process through artificial intelligence technology, including but not limited to the operation steps, processed data, and logical architecture included in the creation process of the AI-generated content. For example, what processing steps are included in the generation of an AI image, and what data are processed in each step. It represents the creation process of the AI-generated content. Different AI-generated content will have different creation processes. Therefore, the generation process information can be used to accurately represent the uniqueness of the AI-generated content.

[0092] Optionally, taking AI images as an example, the AI ​​generation process information may include the following:

[0093] 1. Generate an algorithm or model name: such as the name of the AI ​​model used and the model version number.

[0094] 2. Generation date and time: For example, the specific date and time when the AI ​​image was generated.

[0095] 3. Input parameters or conditions: Input data or conditions (such as text description, sketch, initial image, etc.), any specific configuration parameters (such as image resolution, style options, etc.).

[0096] 4. Detailed information of the generation process: the steps and methods of the generation process, the hardware and software environment used (such as GPU model, operating system, framework version, etc., description of the training dataset, training time and resource consumption, etc.

[0097] 5. Quality assessment and verification information: such as quality assessment indicators, verification and test results of generated images.

[0098] 6. Other relevant information: such as the purpose or application scenario of the generated image, any relevant research papers or document links.

[0099] In one embodiment, step 302 may specifically include: parsing metadata of AI-generated content, and determining generation process information based on the metadata.

[0100] In this embodiment, metadata of AI-generated content refers to additional information related to AI-generated content, which can help understand, manage and optimize the AI ​​generation process. Metadata types include but are not limited to one or more of the generation timestamp, generation algorithm, input data, parameter settings, generation environment, generator information, quality assessment indicators and audit logs of AI-generated content, where the functions of each metadata can be as follows:

[0101] 1. Generation timestamp: used to record the specific time when AI generated content, which helps to track the generation time and version control of content.

[0102] 2. Generation algorithm: used to identify the AI ​​algorithm or model version of AI-generated content, making it easier to understand the technical background of the generated content and reproduce the generation process.

[0103] 3. Input data: Record the input data or initial conditions used for AI-generated content to help analyze the source and generation logic of AI-generated content.

[0104] 4. Parameter settings: including parameter values ​​used in the AI ​​generation process, such as temperature parameters, number of iterations, etc., to facilitate optimization and adjustment of generation strategies.

[0105] 5. Generation environment: Describes the hardware and software environment when AI-generated content is generated, such as operating system version, hardware configuration, dependent library version, etc., to ensure the reproducibility of the generation process.

[0106] 6. Generator information: This records the identification information of the AI ​​system or user that generated the AI ​​content, to facilitate responsibility tracing and authority management.

[0107] 7. Quality assessment indicators: including quality assessment results of AI-generated content, such as accuracy, fluency, relevance scores, etc., to help evaluate the quality of AI-generated content.

[0108] 8. Audit log: records key operations and events in the AI ​​generation process, such as model calls, data processing steps, etc., to enhance the transparency and traceability of the generation process.

[0109] By parsing these metadata, we can fully understand the generation process of AI-generated content and ensure the reliability, traceability and optimization potential of the generated content.

[0110] By parsing the metadata of AI-generated content, key information in the generation process can be effectively extracted and determined. This process can provide a deep understanding of AI-generated content, including information such as the source of the generated content, the parameter settings of the generation algorithm, the generation time, and the generation environment. Through this generation process information, the reliability and quality of AI-generated content can be better evaluated, the generation process can be tracked, and the transparency and traceability of AI content generation can be enhanced.

[0111] In an optional embodiment, taking an AI image as an example, the metadata of the AI ​​image may include the following information:

[0112] 1. Basic information: such as the AI ​​image’s file name, file type, file size, creation date and time, and modification date and time.

[0113] 2. Technical information: such as the resolution, image size, color mode, compression type and level of the AI ​​image.

[0114] 3. Descriptive information: such as the title, description, keywords or tags, classification and other information of the AI ​​image.

[0115] 4. AI generation information: such as the AI ​​image generation algorithm or model name, generation date and time, input parameters or conditions, and detailed information of the generation process.

[0116] In addition, other information of the AI ​​image can be included, such as version history, editing history, related file links, etc.

[0117] These metadata are usually embedded in AI-generated content in a standard format (such as EXIF ​​format). Therefore, the metadata can be parsed by parsing the EXIF ​​file of the AI ​​image to obtain information about the generation process of the AI ​​image.

[0118] In one embodiment, step 302 may specifically include: inputting AI-generated content into a preset recognition model so that the preset recognition model outputs generation process information of the AI-generated content. The preset recognition model is obtained by training a neural network model based on a first sample set, and the first sample set includes first sample content generated based on AI, and the first sample content is marked with corresponding generation process information.

[0119] In this step, the generation process information can be extracted by using a preset recognition model based on a neural network. By inputting the AI-generated content into the preset recognition model, the generation process information of the AI-generated content can be effectively obtained. The key to this process lies in the design and training of the preset recognition model. Specifically, the preset recognition model is obtained by training the neural network model based on the first sample set. The first sample set includes the first sample content generated based on AI, and these sample contents are marked with the corresponding generation process information. The preset recognition model is trained based on samples marked with generation process information, which can converge more quickly and improve the recognition accuracy of the model. Through the recognition model, the specific generation process of AI-generated content can be quickly and accurately tracked and understood, which is of great significance for verifying the authenticity and reliability of the content.

[0120] Step 303: Generate watermark information corresponding to the AI ​​generated content according to the generation process information.

[0121] In this step, watermark information refers to digital watermark information, where the watermark information can be a visible watermark or a dark watermark. Taking dark watermark as an example, dark watermark is a kind of information embedded in digital media (such as images, videos, audio) to protect copyright, verify authenticity or transmit other secret information. Dark watermark is invisible or imperceptible when viewed or used normally, but can be extracted and verified through specific algorithms or tools. The generation process information of AI-generated content can characterize the uniqueness of AI-generated content. Therefore, determining the dark watermark information of AI-generated content based on the generation process information can make the dark watermark uniquely identify the corresponding AI-generated content, thereby realizing the unique identification of AI-generated content without changing the visual or auditory effects of AI-generated content, and improving the traceability and security of AI-generated content.

[0122] In one embodiment, step 303 may specifically include: converting the information format of the generation process information into an information format specified by the target system, and generating watermark information corresponding to the AI ​​generated content according to the generation process information after the format conversion.

[0123] In this embodiment, the target system may refer to a system for managing AI-generated content, such as an e-commerce platform for managing users' AI product information in an e-commerce scenario. By converting the information format of the generation process information to the information format specified by the target system, such as the conversion of the programming language, it can be ensured that the generation process information can be correctly identified and processed in the target system. This format conversion not only improves the compatibility of information, but also reduces the risk of information loss or misunderstanding due to format mismatch.

[0124] Furthermore, by generating watermark information corresponding to the AI-generated content based on the format-converted generation process information, the generation process information can be effectively embedded into the AI-generated content. The advantage of this is that even when the AI-generated content is being disseminated or shared, the generation process information can still be retained and identified, thereby providing a reliable basis for the traceability and verification of the AI-generated content. This way of embedding watermark information is concealed and difficult to tamper with, which enhances the security and credibility of the content.

[0125] In one embodiment, step 303 may specifically include: extracting feature codes of the generation process information, the feature codes are used to characterize the uniqueness of the AI-generated content, and generating watermark information corresponding to the AI-generated content according to the feature codes.

[0126] In this embodiment, the generation process information of AI-generated content may contain a relatively large amount of data. In the process of generating watermark information, the generation process information can be feature extracted. This feature coding can capture unique information in the generation process. Therefore, the feature coding can still be used to characterize the uniqueness of AI-generated content, so that each AI-generated content has a unique identifier. Generating corresponding watermark information based on these feature codes can not only further enhance the copyright protection and traceability of AI-generated content, but also reduce the amount of watermark data calculation and improve calculation efficiency.

[0127] Step 304: embed the watermark information into the AI ​​generated content to obtain the reset target AI generated content.

[0128] In this step, watermark information is embedded into AI-generated content, which does not affect the normal use and display of the content, but can be extracted and verified when needed to ensure the originality and legality of the content. This method not only improves the security and credibility of AI-generated content, but also provides technical support for copyright protection, effectively preventing the content from being illegally copied and tampered with. Improve the traceability and security of AI-generated content.

[0129] In one embodiment, step 304 may specifically include: determining an embedding position of the watermark information in the AI ​​generated content. Determining an embedding strength of the watermark information according to the AI ​​generated content, where the embedding strength is less than or equal to a preset strength threshold. Embedding the watermark information into the AI ​​generated content according to the embedding strength to obtain a reset target AI generated content.

[0130] In this embodiment, embedding strength refers to the strength or amplitude of the watermark signal when the watermark is embedded in AI-generated digital media (such as images, audio or video). It determines the degree of influence of the watermark on the original media. Specifically, the embedding strength can be the mixing ratio of the watermark signal to the original signal. When the embedding strength is low, the watermark signal has less impact on the original media. This means that the watermark is not visually or auditorily noticeable, and the quality of the original media is almost unchanged. When the embedding strength is high, the watermark signal has a greater impact on the original media. This may cause visual or auditory changes, making the watermark more obvious. Therefore, the impact of the watermark on the sensory effect of AI-generated content can be reduced by controlling the embedding strength of the watermark.

[0131] By embedding watermarks in AI-generated content, copyright protection and source tracking of AI-generated content can be effectively achieved. Specifically, the embedding position of the watermark information is first determined. The watermark information can be embedded in a place that is not easily perceived by the human eye. For example, in an AI image, a dark watermark can be embedded in a high-frequency area or an area with more noise. The subtle changes in these areas are not easily perceived by the human eye, so that the dark watermark can be embedded in the AI-generated content without significantly affecting the quality of the original content. Next, according to the characteristics of the AI-generated content, the embedding strength of the watermark information is determined. In order not to significantly change the appearance or sound of the original media, the embedding strength of the watermark information needs to be relatively low. By controlling the embedding strength, the impact on the original media can be minimized while maintaining the detectability of the watermark. For example, the embedding strength can be controlled to be less than or equal to a preset strength threshold, so as to ensure that the watermark information will not have a significant impact on the visual or functional effects of the content after embedding. The preset strength threshold can be adjusted according to actual needs. Finally, the watermark information is embedded into the AI-generated content according to the determined embedding strength to obtain the reset target AI-generated content. This method can not only embed watermarks without significantly affecting the user experience, but also flexibly adjust the embedding strength according to the different characteristics of AI-generated content to ensure the concealment and stability of the watermark. At the same time, the watermark information embedded in this way can play an important role in subsequent copyright verification and infringement tracking, effectively protecting the rights and interests of content creators.

[0132] In an optional embodiment, the embedding strength can be adaptively adjusted according to the characteristics of the AI ​​generated content. For example, a higher embedding strength can be used in high-frequency areas or areas with more noise in the AI ​​image, while a lower embedding strength can be used in low-frequency areas or smooth areas. This approach can minimize the impact on the original AI image while ensuring the robustness of the watermark.

[0133] In an optional embodiment, when the AI-generated content is image content such as images and videos, the perceptual model technology can be used to determine the embedding position. For example, the perceptual model of the human visual system is used to design a dark watermark embedding algorithm. By taking advantage of the difference in sensitivity of the human eye to changes in brightness, color, and contrast, the dark watermark can be embedded in an imperceptible place.

[0134] In an optional embodiment, watermark information can be embedded in AI-generated content using frequency domain transformation, such as using frequency domain transformation (such as discrete cosine transform or discrete wavelet transform) to embed dark watermarks. These techniques convert the media to the frequency domain, embed the watermark in the frequency domain, and then convert back to the time domain. Since the human eye is not sensitive to certain frequency changes, this method can effectively hide the watermark information.

[0135] In an optional embodiment, redundant information can be used to embed watermark information. In actual scenarios, AI-generated digital media usually contains a large amount of redundant information. This redundant information can be used to embed watermarks without significantly changing the perceived quality of the original media.

[0136] For example, the least significant bit (LSB) method can be used to embed dark watermarks, which can embed secret information in digital media such as AI-generated images, audio or video. Watermark information is embedded by modifying the least significant bits of AI-generated content, because these modifications have the least impact on the original data, making the watermark difficult to detect.

[0137] In an optional embodiment, if the AI-generated content has an original work identifier, the original work identifier of the AI-generated content can be directly deleted, and the newly generated watermark information is embedded in the AI-generated content as a new work identifier to ensure that the embedding of the new work identifier does not affect the user experience.

[0138] In one embodiment, before step 303, the method may further include: verifying the data integrity of the AI-generated content. And / or verifying whether the AI-generated content has been tampered with. When it is determined that the AI-generated content is complete and the AI-generated content has not been tampered with, a step of generating watermark information corresponding to the AI-generated content according to the generation process information is performed.

[0139] In this embodiment, before generating watermark information, the steps of data integrity check and / or tampering check can be performed on the AI-generated content, which can significantly improve the security and reliability of the AI-generated content. Specifically, by verifying the data integrity of the AI-generated content, it can be ensured that no data loss or damage occurs during the generation, transmission and storage of the content. By verifying whether the AI-generated content has been tampered with, the content can be prevented from being maliciously modified or tampered with. Only when it is determined that the AI-generated content is complete and has not been tampered with, the step of generating watermark information is performed to ensure the accuracy and effectiveness of the watermark information. The generation of watermark information depends on credible content data, thus providing a reliable basis for subsequent content verification, copyright protection and tracking. This multi-level verification mechanism not only improves the security of the system and the credibility of the content, but also enhances the user's trust in the AI-generated content, ensuring that the AI-generated content is original and untampered. While ensuring content security, it also provides strong support for content management and copyright protection.

[0140] Optionally, for AI-generated content from untrusted systems, a parsing algorithm can be used to extract necessary data and original work identifiers from the AI-generated content. Based on the extracted data and original work identifiers, a preset verification algorithm is used to detect whether the AI-generated content is complete and whether it has been tampered with during transmission.

[0141] In one embodiment, verifying the data integrity of the AI-generated content includes: determining whether each field in the preset field set has matching information in the generation process information. If each field in the preset field set has matching information in the generation process information, it is determined that the AI-generated content is complete. If there is a field in the preset field set that has no matching information in the generation process information, it is determined that the AI-generated content is incomplete.

[0142] In this embodiment, the preset field set includes the types of fields that must be possessed by the complete AI-generated content, and the preset field set can be set in advance according to actual needs. The data integrity of the AI-generated content can be verified by comparing whether the fields are complete. For example, the preset field set can be stored in the form of a field table, and the generation process information is matched with the field table. If the generation process information has no information in some fields in the field table, such as the generation process information of the AI-generated content, the information of field 1 in the field table is empty, indicating that the AI-generated content is incomplete. At this time, a prompt can be issued to facilitate timely processing. If there is matching information about all fields in the field table in the generation process information, it means that the AI-generated content is complete, and the subsequent steps can be automatically executed at this time. By matching each field in the preset field set with the generation process information, it is ensured that all preset fields are included in the generation process information, thereby verifying the integrity of the generated content. When all preset fields have matching information, it can be determined that the AI-generated content is complete, ensuring the reliability and accuracy of the generated content. On the contrary, when there is a field without matching information, the incompleteness of the generated content can be discovered in time, which is convenient for further processing and correction. This verification mechanism improves the quality control of AI-generated content and reduces errors and risks caused by incomplete data.

[0143] In one embodiment, verifying whether AI-generated content has been tampered with includes: using a preset detection model to identify whether the AI-generated content has been tampered with, wherein the preset detection model is obtained by training a neural network model based on a second sample set, the second sample set includes a second sample content pair generated based on AI, and the second sample content pair includes original sample content that has not been tampered with and sample content that has been tampered with.

[0144] In this embodiment, a preset detection model trained based on the second sample set is used to identify tampering of AI-generated content. The second sample set includes original sample content that has not been tampered with and sample content that has been tampered with, which enables the detection model to learn and identify tampering features, thereby effectively detecting tampered content in practical applications. It is possible to efficiently and accurately verify whether the AI-generated content has been tampered with. Since the detection model is trained based on a large amount of sample data, it can accurately identify tampered content, thereby improving the accuracy of tampering detection. The automated detection model can quickly process a large amount of AI-generated content, significantly improve detection efficiency, and save time and cost for manual review. This method can promptly detect and report tampered content, ensure the authenticity and reliability of AI-generated content, and reduce potential risks and losses caused by content tampering.

[0145] In an optional embodiment, by using a diverse second sample set for training, the detection model can adapt to different types and forms of tampering, thereby improving the robustness and adaptability of the system.

[0146] In an optional embodiment, after step 304, the method may further include: generating content configuration qualification information for the target AI according to the watermark information.

[0147] In this embodiment, the reset target AI generated content can be trusted and authenticated through the authentication mechanism of the trust system to verify the uniqueness and integrity of the new work identifier. For example, an authentication request can be sent to the trust system, and the trust system performs authentication processing on the AI ​​generated content and the new work identifier, and feeds back the authentication result, configuring the qualification information of the target AI generated content as "trusted" to facilitate subsequent data management.

[0148] The above-mentioned method for resetting AI-generated content extracts the generation process information of the AI-generated content and generates corresponding watermark information based on the generation process information. Since the generation process information can characterize the uniqueness of the AI-generated content, the corresponding watermark information can uniquely identify the AI-generated content. After the watermark information is embedded in the AI-generated content, the reset target AI-generated content can have an accurate unique identification, thereby resetting the AI-generated content without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content.

[0149] Please see Figure 4 , which is an AI generated content processing method of an embodiment of the present application, the method can be Figure 1 The electronic device 1 shown is used to perform and can be applied to Figure 2 In the reset application scenario of AI-generated content shown in , the AI-generated content can be reset without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content. This embodiment takes dark watermark as an example, and the method includes the following steps:

[0150] Step 401: The user uploads or creates AI-generated content online. For AI-generated content created by the user online, the process may directly proceed to step 412. For AI-generated content uploaded by the user, the process proceeds to step 402.

[0151] Step 402: For AI-generated content uploaded by users, determine the source information of the AI-generated content, and determine whether the AI-generated content comes from a trusted system based on the source information. If yes, proceed directly to step 412. If not, proceed to step 403.

[0152] Step 403: If the AI-generated content does not come from a trusted system, trigger a reset instruction for the AI-generated content.

[0153] Step 404: In response to the reset instruction, obtain the artificial intelligence AI generated content to be processed.

[0154] Step 405: Input the AI-generated content into a preset recognition model so that the preset recognition model outputs generation process information of the AI-generated content.

[0155] Step 406: Verify the data integrity of the AI-generated content. And / or, verify whether the AI-generated content has been tampered with. If the data of the AI-generated content is incomplete and / or the AI-generated content has been tampered with, a prompt message may be issued to facilitate timely processing of illegal data content. When it is determined that the AI-generated content is complete and has not been tampered with, proceed to step 407.

[0156] Step 407: When it is determined that the AI-generated content is complete and has not been tampered with, extract the feature code of the generation process information, where the feature code is used to characterize the uniqueness of the AI-generated content.

[0157] Step 408: Generate dark watermark information corresponding to the AI ​​generated content according to the feature coding.

[0158] Step 409: Determine the embedding position of the dark watermark in the AI ​​generated content.

[0159] Step 410: Determine the embedding strength of the dark watermark information according to the AI ​​generated content, where the embedding strength is less than or equal to a preset strength threshold.

[0160] Step 411: embed the dark watermark information into the AI ​​generated content according to the embedding strength to obtain the reset target AI generated content.

[0161] Step 412: Configure the qualification information of the current AI-generated content as “trusted” to indicate that the current AI-generated content is safe.

[0162] In the above-mentioned AI-generated content processing method, users can choose multiple platforms to produce AI-generated content, or use the system-recommended platform for AI creation. The system identifies whether the source system of the AI-generated content is a trusted system. If it is a trusted system, the qualification information of the current AI-generated content can be directly configured as "trusted" and written into the system. If it is an untrusted system, the AI-generated content is reset, and the qualification information of the reset AI-generated content is configured as "trusted" and written into the system. The embodiment of the present application provides a tool that can re-evaluate and authenticate untrusted AI-generated content, effectively solve the information security issues of AI-generated content from untrusted systems, improve the credibility of AI-generated content, promote the diverse and compliant development of different AI generation models, ensure the legality, credibility and traceability of AI-generated content, and improve the security and reliability of the entire system. And it realizes a new link for identifying, verifying, resetting and re-authenticating untrusted AI-generated content processes, and uses automated workflows and triggers to automatically identify, reset and authenticate untrusted AI-generated content to improve work efficiency.

[0163] For details of each step of the above method, please refer to the relevant description of the above embodiment, which will not be repeated here.

[0164] Please see Figure 5 , which is a method for resetting AI-generated content for a product in one embodiment of the present application. The method can be Figure 1 The electronic device 1 shown is used to perform and can be applied to Figure 2 In the reset application scenario of AI-generated content shown in , this embodiment takes terminal 220 as an execution end as an example. Compared with the previous embodiment, this embodiment takes the reset scenario of AI-generated content about goods in an e-commerce scenario as an example. The method includes the following steps:

[0165] Step 501: In response to a reset instruction, obtain artificial intelligence (AI) generated content to be processed, where the AI ​​generated content includes product information.

[0166] The product information here includes but is not limited to product-related knowledge information, appearance information, product promotion information, product-related order information, product-related service information, etc. For example, the AI ​​main picture of the product, the AI ​​details page of the product, etc. The user command can be a reset command for the uploaded AI-generated content, or it can be a reset command triggered when the user saves the AI-generated content after creating it online.

[0167] Step 502: Extracting the generation process information of the AI ​​generated content.

[0168] Step 503: Generate watermark information corresponding to the AI ​​generated content according to the generation process information.

[0169] Step 504: embed the watermark information into the AI ​​generated content to obtain the reset target AI generated content.

[0170] The above-mentioned method for resetting AI-generated content of commodities realizes the reset processing of AI-generated content in e-commerce scenarios, and realizes the reset of AI-generated content without changing the visual or auditory effects of AI-generated content, thereby improving the traceability and security of AI-generated content. It not only improves the credibility of AI-generated content, but also promotes the diversified and compliant development of various AI-generated models, and improves the information security of AI-generated content in e-commerce scenarios.

[0171] For details of each step of the above method, please refer to the relevant description of the above embodiment, which will not be repeated here.

[0172] Please see Figure 6 , which is a device 600 for resetting AI generated content according to an embodiment of the present application, which can be applied to Figure 1 The electronic device 1 shown can be applied to Figure 2 In the reset application scenario of AI-generated content shown in , the AI-generated content can be reset without changing the visual or auditory effects of the AI-generated content, thereby improving the traceability and security of the AI-generated content. The device includes: an acquisition module 601, an extraction module 602, a generation module 603 and an embedding module 604, and the functional principles of each module are as follows:

[0173] The acquisition module 601 is used to obtain artificial intelligence AI generated content to be processed in response to the reset instruction.

[0174] The extraction module 602 is used to extract the generation process information of the AI-generated content.

[0175] The generation module 603 is used to generate watermark information corresponding to the AI ​​generated content according to the generation process information.

[0176] The embedding module 604 is used to embed the watermark information into the AI ​​generated content to obtain the reset target AI generated content.

[0177] In one embodiment, it further includes: a judgment module, which is used to determine the source information of the AI-generated content before obtaining the AI-generated content to be processed in response to the reset instruction. It is judged whether the AI-generated content comes from a trusted system according to the source information. If the AI-generated content does not come from a trusted system, a reset instruction on the AI-generated content is triggered.

[0178] In one embodiment, the extraction module 602 is used to input the AI-generated content into a preset recognition model so that the preset recognition model outputs the generation process information of the AI-generated content. The preset recognition model is obtained by training a neural network model based on a first sample set, and the first sample set includes a first sample content generated based on AI, and the first sample content is marked with the corresponding generation process information.

[0179] In one embodiment, the extraction module 602 is used to parse metadata of AI-generated content and determine generation process information based on the metadata.

[0180] In one embodiment, the generating module 603 is used to convert the information format of the generating process information into the information format specified by the target system, and to generate watermark information corresponding to the AI ​​generated content according to the converted generating process information.

[0181] In one embodiment, the generating module 603 is used to extract the feature code of the generating process information, where the feature code is used to characterize the uniqueness of the AI ​​generated content, and to generate watermark information corresponding to the AI ​​generated content according to the feature code.

[0182] In one embodiment, the embedding module 604 is used to determine the embedding position of the watermark information in the AI ​​generated content. Determine the embedding strength of the watermark information according to the AI ​​generated content, and the embedding strength is less than or equal to a preset strength threshold. Embed the watermark information into the AI ​​generated content according to the embedding strength to obtain the reset target AI generated content.

[0183] In one embodiment, the method further includes: a verification module for verifying the data integrity of the AI-generated content before generating the watermark information corresponding to the AI-generated content according to the generation process information. And / or, verifying whether the AI-generated content has been tampered with. When it is determined that the AI-generated content is complete and the AI-generated content has not been tampered with, the step of generating the watermark information corresponding to the AI-generated content according to the generation process information is executed.

[0184] In one embodiment, the verification module is specifically used to determine whether each field in the preset field set has matching information in the generation process information. If each field in the preset field set has matching information in the generation process information, it is determined that the AI ​​generated content is complete. If there are fields in the preset field set that have no matching information in the generation process information, it is determined that the AI ​​generated content is incomplete.

[0185] In one embodiment, the verification module is specifically used to use a preset detection model to identify whether the AI-generated content has been tampered with, wherein the preset detection model is obtained by training a neural network model based on a second sample set, and the second sample set includes a second sample content pair generated based on AI, and the second sample content pair includes the original sample content that has not been tampered with and the sample content that has been tampered with.

[0186] For a detailed description of the above-mentioned AI generated content resetting device 600, please refer to the description of the relevant method steps in the above-mentioned embodiment. Its implementation principle and technical effects are similar, and will not be repeated here in this embodiment.

[0187] Figure 7 The following is a schematic diagram of the structure of a cloud device 70 provided in an exemplary embodiment of the present application. The cloud device 70 can be used to run the method provided in any of the above embodiments. Figure 7 As shown, the cloud device 70 may include: a memory 704 and at least one processor 705, Figure 7 A processor is taken as an example.

[0188] The memory 704 is used to store computer programs and can be configured to store various other data to support operations on the cloud device 70. The memory 704 can be an object storage service (OSS).

[0189] Memory 704 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0190] The processor 705 is coupled to the memory 704 and is used to execute the computer program in the memory 704 to implement the solution provided by any of the above method embodiments. The specific functions and technical effects that can be achieved are not repeated here.

[0191] Furthermore, if Figure 7 The cloud device also includes: a firewall 701, a load balancer 702, a communication component 706, a power supply component 703 and other components. Figure 7 Only some components are shown schematically, which does not mean that the cloud device only includes Figure 7 Components shown.

[0192] In one embodiment, the above Figure 7The communication component 706 in is configured to facilitate wired or wireless communication between the device where the communication component 706 is located and other devices. The device where the communication component 706 is located can access a wireless network based on a communication standard, such as WiFi, 2G, 3G, 4G, LTE (Long Term Evolution, Long Term Evolution, referred to as LTE), 5G and other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component 706 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 706 also includes a near field communication (Near Field Communication, referred to as NFC) module to facilitate short-range communication. For example, the NFC module can be based on Radio Frequency Identification (Radio Frequency Identification, referred to as RFID) technology, Infrared Data Association (Infrared Data Association, referred to as IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology and other technologies.

[0193] In one embodiment, the above Figure 7 The power supply component 703 provides power to various components of the device where the power supply component 703 is located. The power supply component 703 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the device where the power supply component is located.

[0194] An embodiment of the present application further provides a computer-readable storage medium, in which computer-executable instructions are stored. When a processor executes the computer-executable instructions, the method of any of the aforementioned embodiments is implemented.

[0195] An embodiment of the present application also provides a computer program product, including a computer program, which implements the method of any of the aforementioned embodiments when executed by a processor.

[0196] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0197] The above-mentioned integrated module implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, including a number of instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor to perform some steps of the methods of various embodiments of the present application.

[0198] It should be understood that the above-mentioned processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in the application can be directly embodied as being executed by a hardware processor, or can be executed by a combination of hardware and software modules in the processor. The memory may include a high-speed RAM (Random Access Memory) memory, and may also include non-volatile storage NVM (Nonvolatile memory, NVM for short), such as at least one disk storage, and can also be a USB flash drive, a mobile hard disk, a read-only memory, a disk or an optical disk, etc.

[0199] The above storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The storage medium can be any available medium that can be accessed by a general or special computer.

[0200] An exemplary storage medium is coupled to a processor so that the processor can read information from the storage medium and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and the storage medium can be located in an application specific integrated circuit (ASIC). Of course, the processor and the storage medium can also exist as discrete components in an electronic device or a main control device.

[0201] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, clothing or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, clothing or device. In the absence of more restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, clothing or device including the element.

[0202] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0203] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods of each embodiment of the present application.

[0204] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of user data and other information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0205] The above are only preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made using the contents of the present application specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present application.

Claims

1. A method for resetting AI-generated content, characterized in that: include: In response to the reset instruction, obtaining artificial intelligence AI generated content to be processed; Extracting generation process information of the AI-generated content; Generate watermark information corresponding to the AI ​​generated content according to the generation process information; The watermark information is embedded into the AI ​​generated content to obtain the reset target AI generated content.

2. The method according to claim 1, characterized in that Before the method of obtaining the artificial intelligence AI generated content to be processed in response to the reset instruction, it also includes: Determining source information of the AI-generated content; Determining whether the AI-generated content comes from a trusted system based on the source information; If the AI-generated content does not come from a trusted system, trigger the reset instruction regarding the AI-generated content.

3. The method according to claim 1, characterized in that The extracting the generation process information of the AI-generated content includes: Inputting the AI-generated content into a preset recognition model so that the preset recognition model outputs the generation process information of the AI-generated content; Among them, the preset recognition model is obtained by training a neural network model based on a first sample set, and the first sample set includes first sample content generated based on AI, and the first sample content is marked with corresponding generation process information.

4. The method according to claim 1, characterized in that: The extracting the generation process information of the AI-generated content includes: Parse metadata of the AI-generated content, and determine the generation process information based on the metadata.

5. The method according to claim 1, characterized in that: The generating watermark information corresponding to the AI ​​generated content according to the generating process information includes: Converting the information format of the generated process information into an information format specified by the target system; The watermark information corresponding to the AI ​​generated content is generated according to the generation process information after format conversion.

6. The method according to claim 1 or 5, characterized in that: The generating watermark information corresponding to the AI ​​generated content according to the generating process information includes: Extracting feature codes of the generation process information, where the feature codes are used to characterize the uniqueness of the AI-generated content; Generate watermark information corresponding to the AI ​​generated content according to the feature code.

7. The method according to claim 1, characterized in that The embedding the watermark information into the AI ​​generated content to obtain the reset target AI generated content includes: Determining an embedding position of the watermark information in the AI-generated content; Determine the embedding strength of the watermark information according to the AI ​​generated content, wherein the embedding strength is less than or equal to a preset strength threshold; The watermark information is embedded into the AI ​​generated content according to the embedding strength to obtain the reset target AI generated content.

8. The method according to claim 1, characterized in that: Before generating the watermark information corresponding to the AI ​​generated content according to the generation process information, the method further includes: Verifying the data integrity of the AI-generated content; and / or verifying whether the AI-generated content has been tampered with; When it is determined that the AI ​​generated content is complete and that the AI ​​generated content has not been tampered with, the step of generating watermark information corresponding to the AI ​​generated content according to the generation process information is performed.

9. The method according to claim 8, characterized in that The verifying the data integrity of the AI-generated content includes: Determine whether each field in the preset field set has matching information in the generation process information; If each field in the preset field set has matching information in the generation process information, it is determined that the AI ​​generated content is complete; If there is a field in the preset field set that has no matching information in the generation process information, it is determined that the AI ​​generated content is incomplete.

10. The method according to claim 8, characterized in that The verifying whether the AI ​​generated content has been tampered with includes: A preset detection model is used to identify whether the AI-generated content has been tampered with, wherein the preset detection model is obtained by training a neural network model based on a second sample set, and the second sample set includes a second sample content pair generated based on AI, and the second sample content pair includes the original sample content that has not been tampered with and the sample content that has been tampered with.

11. A method for resetting AI-generated content about a product, characterized in that: include: In response to the reset instruction, obtaining artificial intelligence (AI) generated content to be processed, wherein the AI ​​generated content includes product information; Extracting generation process information of the AI-generated content; Generate watermark information corresponding to the AI ​​generated content according to the generation process information; The watermark information is embedded into the AI ​​generated content to obtain the reset target AI generated content.

12. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively coupled to the at least one processor; The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the electronic device to perform the method described in any one of claims 1-11.

13. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, and when the processor executes the computer-executable instructions, the method according to any one of claims 1 to 11 is implemented.

14. A computer program product, characterized in that The method comprises a computer program, which implements the method according to any one of claims 1 to 11 when the computer program is executed by a processor.

Images