China: CNIPA Issued Interpretation of the Revision of the Patent Examination Guidelines
Published 7 December 2025
Yu Du
On 10 November 2025, the China National Intellectual Property Administration (CNIPA) issued the revision of the Patent Examination Guidelines, which will take effect on 1 January 2026. The most recent official interpretation, released on 4 December 2025, further clarified the policy objectives and institutional logic of the revision.
This revision focuses on innovation practices in new fields and emerging industries, particularly in high-growth areas such as artificial intelligence, big data, bitstream processing, and plant breeding. In these fields, challenges such as difficulty in defining claims, the mixing of technical contributions and algorithmic rules, and insufficient disclosure have long led to inconsistent examination standards. The revision systematically updates requirements on patentability, inventiveness, and disclosure, while optimizing procedural rules and institutional arrangements in examination practice, improving the precision of patentable subject matter through expanded chapter titles, illustrative examples, and clarified drafting requirements. The key points of CNIPA’s interpretation are as follows:
Adjustments to AI-related Patents
The revised guidelines place a special emphasis on AI-related patents, systematically clarifying, for the first time, the boundaries of ethics and public interest in examination. According to Article 5 of the Patent Law, inventions involving data collection, annotation, training, recommendation, or decision-making that violate laws, regulations, ethical principles, or public interest will be excluded from patentable scope. This provision not only clarifies the legal red line but also provides operational guidance for AI companies in research and development and patent layout.
At the same time, in view of the “black box” characteristics of AI models, the guidelines significantly raise the disclosure requirements in the specification. Contents involving model structure, training methods, parameter selection, and dataset usage must be disclosed in a structured manner, so that technical personnel in the relevant field can actually implement the invention. This measure is intended to prevent enterprises from obtaining improper patent monopolies through opaque algorithms while guiding the industry to form verifiable and trustworthy innovation paths. The interpretation also provides typical examples, such as disclosure norms for model structures and training processes for image recognition, natural language processing, and prediction algorithms, enhancing operational feasibility and consistency in examination.
Bitstream Technology and Digital Content Protection
With the rapid iteration of streaming media, video coding/decoding, and digital content transmission technologies, the demand for patent protection has become increasingly urgent. The revised guidelines, for the first time, explicitly include bitstream inventions within the scope of examination, clarifying requirements for technical feature disclosure and claim drafting. By specifying description requirements for key aspects such as bitstream processing methods, coding algorithms, and data transmission optimization technologies, the guidelines ensure that patent applications are implementable while meeting the standards of inventiveness and sufficient disclosure. This change provides legal certainty for enterprises’ patent layouts in the digital content field and helps avoid invalidation risks due to insufficient description.
Plant Variety Protection and Patent Connection
In the field of plant breeding and biotechnology, the revised guidelines further clarify the scope of patentable plant varieties and strengthen the connection between the patent system and the plant variety protection system. The updated content covers gene editing, molecular breeding, and phenotype improvement technologies, specifying protectable technical features and examination requirements. This not only helps standardize patent examination practices for plant breeding technologies but also provides clear rights boundaries for enterprises in plant breeding, gene editing, and related biotechnologies, reducing the risk of duplicate applications and rights conflicts.
Optimization of Inventiveness and Same-day Application Rules
In substantive examination, the revised guidelines systematically calibrate the assessment of inventiveness and rules for same-day applications. They clearly stipulate that in same-day applications, the invention must forgo utility model rights, which helps reduce disputes arising from “double application–double grant” scenarios and enhances the stability of patent rights. Inventiveness assessment standards are further refined, emphasizing that non-technical features or content not substantially related to technical problems should not be considered as inventive contributions. This has a corrective effect on the examination of inventions such as algorithms, data processing, and business methods, making the correspondence between technical contribution and claim scope clearer and preventing enterprises from packaging “pseudo-technical” patents with abstract data processing or business rules.
Disclosure Requirements and Technical Implementability
The revised guidelines significantly raise the disclosure requirements in specifications, particularly for AI, bioinformatics, and complex algorithm fields. Specifications must ensure that technical personnel in the relevant field can implement the solutions described in the claims. Otherwise, even if the claims are formally complete, they will not meet the sufficient disclosure requirements under Article 26 of the Patent Law. This adjustment strengthens the core principle of the patent system, “disclosure in exchange for protection”, ensuring both the verifiability of patented technologies and prompting innovators to prepare implementable technical solutions prior to filing, thereby reducing invalidation risks.
Procedural Rules and Examination System Improvement
Regarding procedural rules, the revised guidelines clarify standards for fee payment and refund, invalidation and review procedures, and handling of “applicant-caused delays” in patent term compensation, reducing procedural uncertainty for applicants. Furthermore, strengthening the verification of applicant identity, disclosure obligations of inventors, and responsibility of agents in examination helps reduce false filings and issues with nominal inventors. Additionally, practices such as entering the national phase, divisional strategies, and on-demand examination have been institutionalized, aligning the examination system more closely with international practices, improving operational consistency and efficiency, and providing predictable operational standards for examiners and applicants.
Comment
This revision not only implements technical adjustments to existing systems but also represents a systematic upgrade aimed at the future technology ecosystem. The revision closely integrates patent system boundaries with ethics, public interest, and social responsibility, achieving a balance between technological promotion and risk governance, especially in the AI field. By integrating new technology scenarios and refining standards for inventiveness and disclosure, it strengthens a technology contribution-oriented approach and prevents decline in patent quality. Meanwhile, procedural improvements enhance the stability and transparency of examination order, providing stronger institutional certainty for enterprise R&D investment. Overall, the revised guidelines are expected to have a profound impact on the innovation landscape in key fields such as AI, big data, digital content, and biological breeding.
This revision focuses on innovation practices in new fields and emerging industries, particularly in high-growth areas such as artificial intelligence, big data, bitstream processing, and plant breeding. In these fields, challenges such as difficulty in defining claims, the mixing of technical contributions and algorithmic rules, and insufficient disclosure have long led to inconsistent examination standards. The revision systematically updates requirements on patentability, inventiveness, and disclosure, while optimizing procedural rules and institutional arrangements in examination practice, improving the precision of patentable subject matter through expanded chapter titles, illustrative examples, and clarified drafting requirements. The key points of CNIPA’s interpretation are as follows:
Adjustments to AI-related Patents
The revised guidelines place a special emphasis on AI-related patents, systematically clarifying, for the first time, the boundaries of ethics and public interest in examination. According to Article 5 of the Patent Law, inventions involving data collection, annotation, training, recommendation, or decision-making that violate laws, regulations, ethical principles, or public interest will be excluded from patentable scope. This provision not only clarifies the legal red line but also provides operational guidance for AI companies in research and development and patent layout.
At the same time, in view of the “black box” characteristics of AI models, the guidelines significantly raise the disclosure requirements in the specification. Contents involving model structure, training methods, parameter selection, and dataset usage must be disclosed in a structured manner, so that technical personnel in the relevant field can actually implement the invention. This measure is intended to prevent enterprises from obtaining improper patent monopolies through opaque algorithms while guiding the industry to form verifiable and trustworthy innovation paths. The interpretation also provides typical examples, such as disclosure norms for model structures and training processes for image recognition, natural language processing, and prediction algorithms, enhancing operational feasibility and consistency in examination.
Bitstream Technology and Digital Content Protection
With the rapid iteration of streaming media, video coding/decoding, and digital content transmission technologies, the demand for patent protection has become increasingly urgent. The revised guidelines, for the first time, explicitly include bitstream inventions within the scope of examination, clarifying requirements for technical feature disclosure and claim drafting. By specifying description requirements for key aspects such as bitstream processing methods, coding algorithms, and data transmission optimization technologies, the guidelines ensure that patent applications are implementable while meeting the standards of inventiveness and sufficient disclosure. This change provides legal certainty for enterprises’ patent layouts in the digital content field and helps avoid invalidation risks due to insufficient description.
Plant Variety Protection and Patent Connection
In the field of plant breeding and biotechnology, the revised guidelines further clarify the scope of patentable plant varieties and strengthen the connection between the patent system and the plant variety protection system. The updated content covers gene editing, molecular breeding, and phenotype improvement technologies, specifying protectable technical features and examination requirements. This not only helps standardize patent examination practices for plant breeding technologies but also provides clear rights boundaries for enterprises in plant breeding, gene editing, and related biotechnologies, reducing the risk of duplicate applications and rights conflicts.
Optimization of Inventiveness and Same-day Application Rules
In substantive examination, the revised guidelines systematically calibrate the assessment of inventiveness and rules for same-day applications. They clearly stipulate that in same-day applications, the invention must forgo utility model rights, which helps reduce disputes arising from “double application–double grant” scenarios and enhances the stability of patent rights. Inventiveness assessment standards are further refined, emphasizing that non-technical features or content not substantially related to technical problems should not be considered as inventive contributions. This has a corrective effect on the examination of inventions such as algorithms, data processing, and business methods, making the correspondence between technical contribution and claim scope clearer and preventing enterprises from packaging “pseudo-technical” patents with abstract data processing or business rules.
Disclosure Requirements and Technical Implementability
The revised guidelines significantly raise the disclosure requirements in specifications, particularly for AI, bioinformatics, and complex algorithm fields. Specifications must ensure that technical personnel in the relevant field can implement the solutions described in the claims. Otherwise, even if the claims are formally complete, they will not meet the sufficient disclosure requirements under Article 26 of the Patent Law. This adjustment strengthens the core principle of the patent system, “disclosure in exchange for protection”, ensuring both the verifiability of patented technologies and prompting innovators to prepare implementable technical solutions prior to filing, thereby reducing invalidation risks.
Procedural Rules and Examination System Improvement
Regarding procedural rules, the revised guidelines clarify standards for fee payment and refund, invalidation and review procedures, and handling of “applicant-caused delays” in patent term compensation, reducing procedural uncertainty for applicants. Furthermore, strengthening the verification of applicant identity, disclosure obligations of inventors, and responsibility of agents in examination helps reduce false filings and issues with nominal inventors. Additionally, practices such as entering the national phase, divisional strategies, and on-demand examination have been institutionalized, aligning the examination system more closely with international practices, improving operational consistency and efficiency, and providing predictable operational standards for examiners and applicants.
Comment
This revision not only implements technical adjustments to existing systems but also represents a systematic upgrade aimed at the future technology ecosystem. The revision closely integrates patent system boundaries with ethics, public interest, and social responsibility, achieving a balance between technological promotion and risk governance, especially in the AI field. By integrating new technology scenarios and refining standards for inventiveness and disclosure, it strengthens a technology contribution-oriented approach and prevents decline in patent quality. Meanwhile, procedural improvements enhance the stability and transparency of examination order, providing stronger institutional certainty for enterprise R&D investment. Overall, the revised guidelines are expected to have a profound impact on the innovation landscape in key fields such as AI, big data, digital content, and biological breeding.