The use of AI in drafting legal documents raises concerns about professional liability

2–3 minutes

Chatbots are already being used to prepare documents in family disputes. Cases in the United States and a debate within the Brazilian Bar Association (OAB) highlight that the adoption of these tools requires verification and analysis by lawyers.

Artificial intelligence is now being used in divorce proceedings by clients themselves. Instead of just answering questions, some people are turning to chatbots to draft documents and develop strategies that are later presented to lawyers or the judiciary.

This phenomenon is more evident in the United States. A survey conducted by the American law firm Cordell & Cordell in 2026 consulted 133 lawyers from the firm itself. Of these, 80% reported receiving correspondence or drafts from clients that they identified as being produced by AI.

This data reflects the experience of professionals from a single firm and does not imply that 80% of clients use chatbots, nor that the authorship of the documents has been confirmed by an independent audit.

US court identifies increase in AI-generated documents

A decision made in Florida on May 27, 2026, illustrates another problem: the ease of producing large volumes of text. In the case of Robert Hinton Avery v. Danielle Beauzil, the appellate court noted that Avery had filed at least 145 motions in the trial court.

One of his responses to the court consisted of 45 pages generated by AI, accompanied by an 863-page appendix. The judges pointed out unfounded arguments and abusive use of the legal process.

The decision required the court to reject any new filings by Avery that were not signed by a licensed attorney in Florida. While the measure did not prohibit the general use of artificial intelligence, it specifically targeted the conduct of this particular litigant based on the history identified by the judges.

The ruling also emphasized that citing real decisions to support an incorrect legal interpretation can be as problematic as citing a non-existent rule. Responsibility for the documents remains with the person who presents them.

Debate at the OAB warns of risks of relying on automated responses

In Brazil, the OAB’s Special Commission on Multiportas Advocacy debated in October the risks of relying on AI responses without verifying their accuracy and appropriateness. Topics included professional autonomy, data protection, and the procedural consequences of improper use.

The position presented at the meeting was that technology can assist legal practice but does not eliminate the lawyer’s responsibility. According to the commission’s president, Eunice Schlieck, professional knowledge and discernment remain indispensable.

The debate addresses two distinct issues: the use of AI to organize work and the automatic adoption of its conclusions. The cases demonstrate the cost of treating a well-written response as a substitute for analyzing the facts, applicable rules, and the limits of each process.

Frequently Asked Questions

  • Does the research represent all of lawyers’ clients?

    No. The survey collects reports from professionals from a single firm.

  • Does the use of AI eliminate professional liability?

    No. In the OAB debate, technology was presented as an aid that requires the lawyer’s knowledge and discernment.