Nanjing Bans Generative AI in Patent Application Drafting to Ensure Authenticity and Compliance
On June 4, 2025, Nanjing banned AI use in drafting patent applications for pre-examination to ensure authenticity. Violations may lead to rejection or legal penalties.

On June 4, 2025, Nanjing’s Intellectual Property Protection Center (NIPPC) announced a ban on the use of artificial intelligence (AI) for drafting patent application documents submitted for pre-examination. This move affects applicants using China’s regional pre-examination system, which allows for initial review and potential fast-tracking at the National Intellectual Property Administration (CNIPA).
The NIPPC identified cases where patent application content was directly generated by AI, prompting this new policy.
Prohibition Details
- No AI-generated content: Patent application documents must be created manually by the applicant or their authorized patent agency. This includes writing, drawing, editing, and organizing based on actual inventions, research results, and supporting materials.
- No AI for R&D evidence: The use of AI to produce research and development materials—such as experimental data, technical reports, or scientific explanations—is strictly prohibited. All R&D documentation must reflect genuine activities, be accurate, verifiable, and traceable.
Consequences of Violations
Applications found to include AI-generated content will face rejection in the pre-examination stage, accompanied by a “Pre-Examination Quality Notice” sent to both the applicant and their agency.
For serious violations, the NIPPC may suspend the offending party’s pre-examination service qualification and add them to a negative management list. In cases that breach specific articles of the Patent Law and its regulations, the matter will be reported to administrative authorities, potentially leading to legal penalties.
Verification Measures
The NIPPC will implement various checks during pre-examination, including:
- Using text detection tools to assess content originality.
- Engaging expert reviewers to judge the logic, professionalism, and validity of documents.
- Requesting detailed explanations or additional evidence from applicants concerning key document content.
This approach aims to ensure patent applications are grounded in authentic work rather than AI-generated content.
For reference, the original notice is available in Chinese here.