Filing AI-assisted briefs without checking them is, in general, a serious offence: Circular 3/2026

The Plenary of the General Council of the Spanish Legal Profession approved in Madrid, on 20 March 2026, Interpretive Circular 3/2026 of the Code of Ethics, on the use of generative artificial intelligence systems and the duty of verification. According to its text, the circular is issued under Article 23 of Organic Law 5/2024, on the Right of Defence.

Its purpose is clear: to clarify the ethical responsibility of the lawyer who files or delivers briefs drawn up with generative AI when they contain errors due to lack of verification and control.

What it interprets

The circular centres on Article 21 of the Code of Ethics and relates it to its Articles 4.1, 10.2.e and 12.A.8, and to Articles 124.g and 125 of the General Statute of the Spanish Legal Profession.

Its conclusions

1. Using AI is lawful. Using generative AI to prepare drafts is a lawful and admissible practice. What is incompatible with professional ethics is the uncritical delegation of essential functions.
2. AI is an auxiliary. It does not replace the professional: it works as support subject to human supervision.
3. Use without control is, in general, a serious offence. Uncritical use or use without sufficient control can breach Articles 4.1 and 21.2 of the Code and Article 47 of the General Statute, and is in general a serious infringement under Article 125.u) of the Statute.
4. The signer answers. The professional who signs is solely responsible for the content. According to the circular, a technical failure of the tool does not remove responsibility.
5. Developers are not ethically liable for misuse of their tools.
6. Technology does not exempt from compliance with ethical rules.
7. When the infringement is complete. On filing or delivering the document with the intention that it produce effects. A timely correction before that moment can avoid it, but not if the correction comes only after a formal request.
8. Grading. The sanction takes into account the degree of negligence and the harmful result. It may fall below a serious offence or, in extreme cases, be classed as very serious under Article 124.g) of the Statute.

Good practices

The circular adds a series of good practices. They are not mandatory, but according to the text itself they are relevant in assessing the professional's diligence:

• Know the tools being used.
• Read the entire result critically, not only part of it.
• Cross-check against reliable sources.
• Use AI only in matters you master.
• Keep internal traceability of AI use.

Who it affects

The circular interprets the Code of Ethics, which applies to all legal professionals, whether they practise alone, in a collective firm or in a company legal department. Its conclusion on the responsibility of the signer therefore also affects firms that divide work among several lawyers: whoever signs the brief answers for all its content, even if another colleague prepared it with the help of a tool.

What changes in practice

Correction has its moment. The seventh conclusion is especially practical: correcting before filing or delivering can avoid the infringement; correcting only after a formal request cannot. Briefs should be reviewed before they go out, not after.

Traceability protects. Keeping a record of how AI was used in a brief makes it possible to prove diligence if a problem arises. It is a simple practice to add to the firm's internal policy.

It fits with the other documents. The circular specifies in ethical terms what the courts were already warning, as we covered in the review of sanctions for false case law, and what the Spanish Legal Profession's AI decalogue proposes with its principle of personal responsibility.

That each firm should have its own rules of use is what we argued in the editorial on the AI policy at the firm, and the circular reinforces that need. More news in Industry news.