Updated on 2026-07-27 GMT+08:00

Data Source Copyright Check

Data source copyright check, a data source security protection product designed by Huawei Cloud, is used for AI model training. It aims to build copyright risk rating models, helping you quickly identify potential copyright compliance risks in AI training datasets, as well as building reliable, secure, and controllable AI applications.

Unauthorized use of copyrighted data for model training may lead to lawsuits, compensation, or even product removal. DSC data source copyright risk check service can be used to build copyright risk rating models based on large model training copyright infringement cases in and outside Chinese mainland, industry lawyers' research and assessment, license classification, license agreement, copyright statement, commercial scenario, and user type. There are millions of built-in copyright risk information, providing accurate copyright risk detection and risk governance capabilities.

Prerequisites

Constraints

Item

Description

Copyright risk check

  • The copyright risk check is performed only on the data source URL level, and does not involve actual scanning or identification of dataset content.
  • Only the existing source copyright information can be used for risk assessment, including the built-in copyright risk information of DSC and the sources in the user-defined copyright library. Currently, copyright risk assessment is not supported for unknown data sources.

Adding data source to the user-defined copyright library

The user-defined copyright library uses the domain name as the unique identifier. The following data sources cannot be added to the user-defined copyright library: data source with an existing domain name and data source in the DSC built-in copyright library.

Specifications

  • A data source file cannot be larger than 20 MB.
  • A task can contain a maximum of 10,000 data source URLs.
  • A tenant can create a maximum of 1,000 tasks in total.
  • A maximum of 10,000 data sources can be configured in the custom copyright library.

Region

This function is supported only in the CN North-Beijing4, CN East-Shanghai1, and CN Southwest-Guiyang1 regions.

  1. Log in to the DSC console.
  2. Click in the upper left corner and select a region or project.
  3. In the navigation pane on the left, choose AI Data Security > Data Copyright Compliance. The Data Copyright Compliance page is displayed.
  4. In the Check Tasks tab, click Download Copyright Source File Template.xlsx under Prepare Metadata File to download the template and configure the data source information as required.
  5. Log in to the OBS console and upload the file created based on the template.
  1. Log in to the DSC console.
  2. Click in the upper left corner and select a region or project.
  3. In the navigation pane on the left, choose AI Data Security > Data Copyright Compliance. The Data Copyright Compliance page is displayed.
  4. In the Check Tasks tab, click Create Task in the upper left corner.
  5. In the displayed dialog box, create a copyright source file template, and enter the bucket name and file name of the uploaded file.
  6. Click Start Check Task and view the task status in the task list.

    Figure 1 Selecting a copyright source file

  7. If Check Status is Completed, the check task is created and completed.
  1. Log in to the DSC console.
  2. Click in the upper left corner and select a region or project.
  3. In the navigation pane on the left, choose AI Data Security > Data Copyright Compliance. The Data Copyright Compliance page is displayed.
  4. In the Check Tasks tab, locate the task and click Confirm in the Operation column.
  5. Select Copyright Risk Level: DSC analyzes the data sources one by one based on the built-in risk rating model, identifies the copyright and license information, and outputs the risk rating result. You can manually adjust the identification results as required.
  6. Confirm the scan results. You can export the data source copyright risk assessment report only after all scan results are confirmed.

    The copyright risk identification result of each data source can be confirmed only once.
    • To batch confirm the results, select them on the left and click Batch Confirm.
    • To confirm all results, click Confirm All in the upper left corner.
    • To confirm a single result, locate it in the list and click Confirm in the Operation column.

You can export the check report to a specified OBS path.

  1. Log in to the DSC console.
  2. Click in the upper left corner and select a region or project.
  3. In the navigation pane on the left, choose AI Data Security > Data Copyright Compliance. The Data Copyright Compliance page is displayed.
  4. In the Check Tasks tab, locate the task and click Export or Export Again in the Operation column. Before exporting the check result, ensure that all check results are confirmed.
  5. In the displayed dialog box, specify the OBS path for saving the check report.
  6. Click Export to export the check report. In the Export Status column, click View Export Result to go to the OBS console and view the check report.
  1. Log in to the DSC console.
  2. Click in the upper left corner and select a region or project.
  3. In the navigation pane on the left, choose AI Data Security > Data Copyright Compliance. The Data Copyright Compliance page is displayed.
  4. In the Copyright Library Maintenance tab, click Batch Add.
  5. In the displayed dialog box, click Download the copyright library batch import template to download the template. Fill the template as required.
  6. Click Add, select the template file configured in Step 5, and click Open.
  7. Click Import to import the template content. During the import, the system checks whether the domain name is the same as that in the DSC built-in risk rating model or an existing custom copyright library. If a conflict is detected, the adding will fail, and the detailed reason will be returned.
  1. Log in to the DSC console.
  2. Click in the upper left corner and select a region or project.
  3. In the navigation pane on the left, choose AI Data Security > Data Copyright Compliance. The Data Copyright Compliance page is displayed.
  4. In the Copyright Library Maintenance tab, click Add Copyright Information in the upper left corner.
  5. Configure the parameters by referring to Table 1. During the import, the system checks whether the domain name is the same as that in the DSC built-in risk rating model or an existing custom copyright library. If a conflict is detected, the adding will fail, and the detailed reason will be returned.

    Table 1 Copyright information

    Parameter

    Description

    Domain

    Enter a unique domain name to identify the data source. For example, nebraska.forums.rivals.com.example.com.

    Copyright License Address (Optional)

    License URL of the data source, which is used for tracing and query. For example, https://example.com/licenses/standard.

    Copyright License Name (Optional)

    Name of the copyright license agreement, which is used to quickly identify the license type. For example, Standard AI Training License v1.0.

    Copyright License Content (Optional)

    Core terms of the license, which specify the usage scope and restrictions.

    Example: The dataset can be used for non-commercial AI training and cannot be redistributed.

    Copyright Address (Optional)

    Original copyright statement or source page link of the dataset. For example, https://example.com/copyright-info.

    Copyright Authorizer (Optional)

    The copyright holder and complete copyright statement, including the year and rights reservation description, facilitate copyright tracing and compliance management. For example, ©2015-2025 Example. All Rights Reserved.

    Copyright Risk Level

    Select a copyright risk level from the drop-down list:

    • Critical
    • High
    • Medium
    • Low
    • Information
    • Unknown

    Copyright Identification Time (Optional)

    Click to set the copyright identification time. Record the copyright identification time by the system or user, which is used for auditing and tracing.

  6. Click OK.

Related Operations

  • To delete a check task, locate it in the Check Tasks tab and click Delete in the Operation column. To delete multiple tasks, select them and click Batch Delete in the upper left corner.
  • Copyright library maintenance:
    • To delete a copyright record, locate it in the Copyright Library Maintenance tab and click Delete in the Operation column. To delete multiple copyright records, select them and click Batch Delete in the upper left corner.
    • To edit a copyright record, locate it in the Copyright Library Maintenance tab and click Edit in the Operation column