China’s latest revision to the Patent Examination Guidelines, published by CNIPA at the end of 2025 and effective from January 1, 2026, makes the examination framework for inventions involving artificial intelligence, big data and algorithm-related features significantly more explicit. The revised text renames the relevant section to cover inventions involving AI, big data and other solutions containing algorithmic features or business rules and methods, while also adding a clearer Article 5(1) filter for content that violates law, social morality or the public interest.
This is more than a drafting clean-up. For applicants, R&D teams and patent firms, the revised framework shifts the focus from simply asking whether a solution “uses AI” to asking who the real inventors are, whether the data and decision logic are legally and ethically defensible, whether algorithmic features and technical features together form a real technical contribution, and whether model design or encoding-related subject matter has been disclosed with enough specificity to support patent protection. In the same revision package, CNIPA also introduced dedicated rules on bitstream-related claims, reinforcing the same policy instinct: abstract data outputs do not automatically deserve protection, but concrete technical methods tied to how such outputs are generated, stored or transmitted may.
Log in to continue reading
Full content is available to registered users only, including detailed analysis and practical recommendations.