Image generation and content detection tasks

Create a task

1. Log on to the AI DeepSign console. In the left-side navigation pane, choose New Task.

2. After you enter the page, the system opens a new task window by default. You can also click the New Task button to create a conversation task.

   Note

   If you have closed the task sidebar, click in the upper-left corner to create a task.

Generate AI images and automatically sign

Use conversational AI to generate images. The system automatically adds C2PA basic signatures to the generated images, ensuring that every AI-generated work has a verifiable unique identity.

1. On the task conversation page of New Task, click the Generate AI Image and Automatically Sign feature card at the bottom. You can also directly enter an image generation request in the dialog box. Example: "Generate a sunset seascape".

2. Press the send key (or press Enter). The system automatically starts generating the image and completes the digital signature.

3. After the image is generated, preview it in the dialog box. In the C2PA Signing Complete area below the preview image, view the signature information, including:

   • Signature Algorithm:the name of the algorithm used for signing.

   • Issuer:the certificate authority (CA) that issued the signing certificate.

   • Timestamp:the time when the signature was generated.

4. Click Download Signed File to download the image with the C2PA digital signature to your on-premises device.

5. (Optional) Click C2PA Official Validation to go to the C2PA official verification site and verify the signature format.

Detect and sign generated AIGC content

Upload existing AI-generated content (AIGC). The system automatically detects whether the content is AI-generated and adds a C2PA basic signature to it.

1. On the task conversation page of New Task, click the upload button in the dialog box and select the image file to detect.

2. After the image is uploaded, enter a detection request in the message input box (for example, "Detect the image"), and press the send key (or press Enter).

3. The system automatically completes detection and analysis and displays the following results in sequence:

   • File Parsing Results: basic information such as file size, format, and resolution.

   • Sensitive information detection: detects whether the image contains sensitive data, which is displayed by risk level (S1 low risk, S2 medium risk, and S3 high risk).

   • Signature certificate tracing: detects whether a C2PA digital signature already exists in the image.

   • AI generation detection: determines whether the image is AI-generated and provides a confidence score (90% or higher: high risk, 60% to 89%: medium risk, and lower than 60%: low risk).

4. (Optional) Add a C2PA digital signature to an unsigned image:

   1. In the Digitally Sign the Content with C2PA? area, click Personal Certificate Signature. The system generates a C2PA basic signature for the image.

   2. After the signature is complete, view the signature details in the C2PA Signing Complete area, and click Download Signed File to download the signed image.

View task history

All generation and detection tasks are automatically saved as historical records, which are displayed in groups by time (today, this week, this month, and more). You can quickly trace task details and the responses of the AI assistant.

1. On the task conversation page of New Task, view the task execution history in the left-side navigation pane.

   Note

   If you have closed the task sidebar, click in the upper-left corner to view the list of historical tasks.

2. Click a historical record to view the task details and signature information:

   Note

   Signing operations cannot be re-executed in historical sessions. To sign content, create a task.

   • Task details: task type (image generation, content detection, and more), execution time and status, responses of the AI assistant, and generated images.

   • C2PA digital signature: for generated content, you can view and download the signed image again.

Generate trusted signatures

When you need to integrate content credential signing into existing content production or management workflows, AI DeepSign supports creating C2PA trusted signature tasks for individual images, videos, or objects in an OSS bucket through OpenAPI.

• Image/video signing: call CreateC2paSignTask to create a signing task for a single JPEG, PNG, or MP4 file. Video signing reuses this API. No dedicated video signing API is provided. The FileUrl parameter must be an absolute HTTP or HTTPS URL of the file to be signed, and can be up to 4,096 characters in length.

• OSS Bucket batch signing: call CreateC2paBucketSignTask to batch process objects under a specified prefix in an OSS bucket.

  • Before you call this API, authorize the service-linked role AliyunServiceRoleForAIDeepSign to access the target bucket.

  • The BucketRegionId parameter must match the actual region of the bucket.

  • The SourcePrefix parameter must be ai-deepsign/ or its subdirectory. The service recursively processes objects with the .jpg, .jpeg, .png, or .mp4 extension under this prefix.

Query task results

Call GetC2paSignResult to query the status and results of a single-file or OSS Bucket batch signing task.