China is strengthening its ethical review process for artificial intelligence patents.
This will ensure that new AI technologies follow the law, support public interests, and reflect social values.
Officials announced the changes at a press conference in Beijing, highlighting the country’s focus on responsible innovation.
New Guidelines
The China National Intellectual Property Administration (CNIPA) will introduce updated patent review guidelines on Jan. 1, 2026.
These guidelines include, for the first time, a standalone section devoted entirely to AI and big data.
According to senior CNIPA official Jiang Tong, this new section sets clear expectations.
It states that any AI-related technical solution, such as data collection, model design, or rule development, must meet legal and ethical standards.
This requirement applies to all applicants, including research institutions, private companies, and foreign-invested firms.
AI’s “Black Box”
Many AI systems operate in ways that are hard to explain. This creates challenges during patent reviews, where clarity and detailed disclosures are essential.
To address this issue, the new guidelines outline how inventors must describe their AI models.
They must explain model construction and training steps in a clear and complete way.
This helps ensure that reviewers understand the invention and can verify that it meets patent disclosure requirements. In turn, it reduces the risk of unclear or incomplete filings.
International Cooperation
China also plans to expand international cooperation on intellectual property over the next five years.
At the same event, CNIPA official Liang Xinxin said the country’s 2026–2030 IP development plan will continue to promote global exchanges.
Moreover, China reaffirmed its commitment to providing equal IP protection for both domestic and foreign-invested enterprises.
This will strengthen business confidence, encourage fair competition, and support the smooth flow of goods and innovation across markets.
