China: the Supreme People’s Court’s 2026 Opinions, the 2022 judicial AI policy, and disclosure at Xiamen

The Supreme People’s Court Opinions, 7 September 2026

• StandingAn Opinions document, Fa Fa [2026] No. 10: national judicial guidance to every people’s court, not legislation. China has no dedicated AI statute; the Opinions work from the Civil Code, the Cybersecurity Law, the Copyright Law and the Civil Procedure Law among others.
• Art. 19Litigation participants submitting AI-generated litigation documents, case search reports and similar materials shall verify the truthfulness and accuracy of the relevant laws, judicial interpretations and cases before filing, explain the AI assistance on filing, and bear legal responsibility for the content. Working translation.
• Art. 19Obtaining false evidence by deleting or altering the generated-content identifier, by specific instruction inputs, selective presentation of results or adversarial interference, and bringing false litigation on it: claim rejected, fines or detention by seriousness, criminal liability where a crime is constituted.
• Art. 18Courts focus review on the truthfulness and integrity of the process by which electronic data was generated, collected, stored and transmitted, with separate review points for big-data reports and blockchain-stored evidence.
• Art. 17Where the party holding documentary or electronic evidence refuses without proper reason to produce it, the court may find the other side’s assertion about its content established.
• Art. 12An AI developer raising a non-infringement defence in a copyright dispute is ordered to produce its training data sources, training process records, mode of model operation and scientific basis.

The full page on the 2026 Opinions, with the Chinese text of the operative sentences beside the working translation.

The Supreme People’s Court Opinions, 2022

• Issued9 December 2022, well before almost every other instrument on this site.
• AimDeep integration of AI with judicial work, deepening the construction of “smart courts”, and a higher level of digital justice.
• DeadlinesAll courts to implement improved AI functional systems by 2025, reaching full application effectiveness by 2030.
• Supported usesEvidence guidance and review; retrieval of laws, regulations and similar cases; adjudication assistance; AI-assisted generation and review of legal documents. The stated purpose is uniformity of adjudication.

Xiamen Maritime Court: disclosure, and a prohibition

In 2025 the Xiamen Maritime Court ruled on a lawyer’s use of AI in civil proceedings, requiring full disclosure, and subsequently issued Guidelines for Litigation Participants’ Use of Artificial Intelligence on a trial basis.

• PermittedLitigants are generally permitted to use AI.
• ProhibitedUsing AI to generate, fabricate, alter or tamper with evidence, or to distort factual circumstances.
• RequiredFull disclosure of AI use.

The evidence prohibition is the same rule Qatar wrote at paragraph 18 of its practice direction in January 2026, and it is stricter than anything currently in force in the United States, where proposed Federal Rule of Evidence 707 was never approved. The 2026 Opinions carry the disclosure duty from one court to all of them.

Read it beside the labelling measures

China now has both halves of a provenance regime. The Measures for Labelling of AI-Generated Synthetic Content, in force since 1 September 2025, require explicit visible labels and implicit machine-readable metadata carrying the provider name and a content reference number. Article 19 of the 2026 Opinions names deleting or altering that identifier as one of the ways false evidence is obtained.

One regime marks the file. The other governs what may be done with it in a courtroom, and now says what happens when the mark is removed.

Sources

• Supreme People’s Court, release of the Opinions on Lawfully Adjudicating Cases Involving AI Disputes, 7 September 2026, full text. Read in full in Chinese; working translation.
• Supreme People’s Court Opinions (2022), China Justice Observer
• Xiamen Maritime Court, Guidelines for Litigation Participants’ Use of Artificial Intelligence (trial), 2025.
• The 2022 Opinions and the Xiamen Guidelines were not read at source. Source: China Justice Observer, Akin and Reed Smith analysis. Those two sections work from translations and carry no quotation and no article numbers.

Related

• AI rules for lawyers, country by country
• China · Supreme People’s Court Opinions on AI disputes, 2026
• China · AI content labelling measures
• China · Generative AI Interim Measures
• Qatar, the same evidence prohibition · Ireland, the same verification duty

Checked 9 September 2026. The 2026 Opinions read at source; the 2022 Opinions and the Xiamen Guidelines worked from translations and analysis. Information, not legal advice, and not advice on the law of the People’s Republic of China.

Common questions

Does China require verification of AI-generated court filings?

Yes, as national judicial guidance. On 7 September 2026 the Supreme People's Court issued Opinions on Lawfully Adjudicating Cases Involving Artificial Intelligence Disputes, Fa Fa [2026] No. 10. Article 19 directs litigation participants to verify the relevant law and cases before submitting AI-generated court documents and case search reports, to explain the AI assistance when submitting them, and to bear legal responsibility for their accuracy. The document is judicial guidance from the highest court, not legislation. The English is Regulayer's working translation.

Does China regulate AI in its courts?

Yes, and earlier than most. On 9 December 2022 the Supreme People's Court issued Opinions on Regulating and Strengthening the Applications of Artificial Intelligence in the Judicial Fields, a programme for all courts to implement AI systems by 2025 and reach full application effectiveness by 2030. The 2026 Opinions add the rules for deciding AI disputes and for AI-generated material in litigation.

Do Chinese courts require lawyers to disclose AI use?

Since 7 September 2026, nationally: Article 19 of the Supreme People's Court Opinions directs litigation participants to explain the AI assistance used when submitting AI-generated litigation documents. Before that, the Xiamen Maritime Court's 2025 Guidelines for Litigation Participants' Use of Artificial Intelligence required full disclosure on a trial basis, and prohibited using AI to generate, fabricate, alter or tamper with evidence.

Is China's approach to AI in courts different from the West's?

In one respect substantially. Most Western instruments regulate AI as a risk to be contained. China's 2022 Opinions are framed as a programme of adoption with deadlines. The 2026 Opinions then reach the same duties Western courts have been writing: verify before filing, disclose on filing, and review electronic evidence for the integrity of the process that produced it.