Research Integrity for Using Artificial Intelligence

Research Ethics Committee of Chinese Academy of Sciences

2024.9.10

To set up a standard for using artificial intelligence in scientific research activities, to avoid risks to research integrity caused by its misuse or abuse, and in adherence to the principles of honesty, transparency, and responsibility, the Research Ethics Committee of Chinese Academy of Sciences issues the following guidelines to our scientific and technical personnel and students, in accordance with the integrity management requirements for the entire research process:

1. During topic selection, literature review, and data collection, artificial intelligence may be used to track research developments and to collect and organize references. However, users should assess the authenticity, accuracy, and reliability of AI-generated information. The direct use of unverified AI-generated research reports, topic suggestions, or literature reviews is discouraged.

2. When preparing application materials, if AI-generated content is used, the user must take full responsibility for the content and provide a complete and truthful disclosure of its use. The direct use of unverified AI-generated application materials is discouraged.

3. When collecting and using data, if AI-generated simulation or test data is used, or if AI technology is applied to perform statistical analysis on raw data, the use of such methods must be fully and truthfully disclosed. AI-generated data must not be presented as experimental data.

4. When producing audio, video, figures and tables, AI technology may be used as an aid. The generated content should be clearly labeled, and the use of AI must be fully and truthfully disclosed. The direct use of AI-generated audio, video, and visual materials is discouraged.

5. When writing research outputs, AI technology may be used to assist in organizing existing theories, materials, and methods, as well as for language polishing, translation, and format checking. However, AI-generated content must not be presented as core innovative results, nor should the entire research outputs or reference lists be generated directly by AI.

6. In peer review, the use of AI technology to write review comments is discouraged, and review information must not be uploaded to any tool or platform that has not been approved by the review organizer.

7. In scientific research, if AI technology is used, the name, version, date, and usage process of the AI should be disclosed in the notes, acknowledgments, references, or appendices. The undisclosed use of such tools is discouraged.

8. When choosing AI technology, platforms and services that have been registered and filed with national authorities should be used. The misuse of AI technology in ways that compromise data security, infringe intellectual property rights, or leak personal privacy is strictly discouraged.

All affiliated institutes of the Chinese Academy of Sciences should, based on the above guidelines, and in consideration of their specific disciplines and academic practices, conduct necessary training to guide scientific and technical personnel and students in using AI technology transparently, compliantly, and responsibly.