top of page

AI Hallucinations or Human Error? Professional Responsibility Hasn't Changed

18 hours ago
2 min read

Updated: 5 minutes ago

AI in Legal Practice: AI Hallucinations or Human Error? Professional Responsibility Hasn't Changed. Marie Odtojan
AI in Legal Practice: AI Hallucinations or Human Error? Professional Responsibility Hasn't Changed

19 November 2026

AI | Law | European Court of Human Rights | Legal Professional Obligations


The European Court of Human Rights' decision to exclude a Romanian lawyer from representing applicants for two years over unverified AI-generated legal citations raises a broader issue about professional competence, diligence and accountability


What surprises me is not that AI can generate fictitious cases or inaccurate information. We already know that AI makes mistakes. What concerns me is that a legal practitioner would submit such material to the court without independent verification.


Verifying legal authorities is standard legal practice. It is not a new obligation introduced by AI.


Lawyers have always been expected to check that the authorities they cite exist, that the references are accurate and that those authorities actually support the legal propositions advanced.


Notably, in this case, the Court identified not only fictitious cases and paragraphs, but also genuine authorities that were irrelevant or contradicted the submissions they were cited to support.


Submitting unverified authorities is no different from pleading a contractual claim without verifying the existence of the contract, its terms and the evidentiary basis for the relief sought, before signing the required certification and filing it with the court.


Why should these standards change simply because AI was used?


What I find ironic is that lawyers warn the public, including self-represented litigants, not to treat AI as a substitute for a qualified lawyer. Yet when lawyers submit unverified AI-generated material, aren't they effectively doing the very thing they warn others against?


AI is not another lawyer. It is a tool.


As they say in IT: Garbage in, garbage out. But it goes beyond the quality of the input. It also concerns the user's competence, critical thinking and ability to verify the output.


In my legal experience, I've witnessed serious deficiencies in professional diligence and verification on a number of occasions that had nothing to do with AI.


This raises another question: If a practitioner relies on unverified AI-generated material in court, what does that suggest about their approach to verification in other aspects of their legal practice?


This also raises an interesting point in the wider debate about AI replacing professionals, including lawyers. AI may become increasingly sophisticated and accurate, but lawyering involves more than generating information. It requires critical analysis, applying the law to the facts, exercising professional judgment and understanding the nuances of each case.


AI can assist with these tasks, but it should not be treated as a substitute for independent professional judgment and accountability.


Technology may evolve, but the need for professional oversight, verification and accountability remains.


Marie Odtojan | Legal Commentary & Opinion

Disclaimer: This article contains legal commentary and opinion for general information and educational purposes only. It does not constitute legal advice.


Comments


Featured Posts
Recent Posts
Search By Tags
  • Facebook Classic
  • LinkedIn App Icon
  • Twitter Classic
  • Google Classic
Follow Us
  • Facebook Metallic
  • LinkedIn Metallic
  • Twitter Metallic

Information on this website is not legal advice.

Liability limited by a scheme approved under Professional Standards Legislation.

 © 2024 All rights reserved for Odtojan Bryl Lawyers.

bottom of page