The rapid growth of artificial intelligence in the legal sector creates many new opportunities. At the same time, it also introduces serious challenges, especially regarding accuracy and professional responsibility. Two recent judgements, Ko v. Li from the Ontario Superior Court of Justice in Canada and a ruling from the Divisional Court in the United Kingdom involving Ayinde v. Haringey and Al-Haroun v. Qatar, serve as strong warnings to legal professionals globally, including those in Vietnam.
In the Canadian case Ko v. Li, lawyer Jisuh Lee was
suspected of using a generative AI tool to prepare a legal factum that
contained several inaccurate or entirely fabricated case citations. The court
found that some of the cited judgements did not exist, while others were quoted
in a way that misrepresented their actual content. Justice F.L. Myers stressed
that lawyers must be honest when citing case law. They have a duty to
personally review and verify all submitted materials, especially those created
with technological assistance. He also warned that submitting false or
misleading citations could be considered contempt of court and required the
lawyer to explain her actions.
In the United Kingdom, under its inherent "Hamid"
jurisdiction, the Divisional Court examined two similar incidents involving the
use of AI in legal submissions. In Ayinde v. Haringey, solicitor Sarah
Forey included several non-existent cases in her argument and misunderstood the
content of a statutory provision. The court found her conduct seriously
negligent, ordered her to pay wasted costs, and referred the matter to the
relevant regulatory authorities. In Al-Haroun v. Qatar, lawyer Abid
Hussain submitted a legal document with eighteen fake case citations out of
forty-five. These references were generated by the client using public AI
tools, and Mr. Hussain did not verify them independently. Although the court
found no evidence of intentional dishonesty, it referred the case to the
Solicitors Regulation Authority for further review.
Both cases highlight several important points. First,
artificial intelligence, especially tools based on large language models,
cannot replace reliable and professional legal research. Second, AI
hallucinations are a known issue. These occur when AI produces content that
sounds reasonable but is actually false. Third, lawyers have a clear ethical
and legal duty to verify all information before using it in their work, even if
that information came from an AI tool. Fourth, professional organisations, such as the Bar Council, the Law Society, and other national bar associations, must go
beyond issuing guidelines. They must also take action to inspect, monitor, and
address misconduct related to AI use.
From these international lessons, Vietnam must take timely
steps to establish a legal and ethical framework for AI in the legal
profession. The use of AI to generate fake citations has already occurred in
Vietnam. Therefore, the following actions should be considered essential.
First, the Vietnam Bar Federation and law schools should
issue clear guidelines about the duty to verify information obtained through AI
tools. This is especially important in litigation, where errors can lead to
serious consequences.
Second, lawyers should be required to disclose whether they
used AI when preparing legal documents. They should identify the tools used,
describe how they verified the information, and accept full responsibility for
the accuracy of the content. Saying that the mistake was caused by AI should
never be used to avoid accountability.
Third, responsible AI use should become an official training program for lawyers, judges, and legal professionals. In
addition to technical skills, education should emphasise ethical values, risk
awareness, and professional standards.
Fourth, there should be a formal process for courts and
legal associations to cooperate in handling violations related to AI use. This
may include administrative penalties, suspension of practice, or in serious
cases, criminal investigation.
The hallucination effect and the black-box nature of artificial intelligence are technical limitations that cannot be entirely avoided. However, in the legal field, where accuracy and integrity are essential, the use of AI must be handled with extreme care. The cases of Ko v. Li, Ayinde v. Haringey, and Al-Haroun v. Qatar are clear warnings. Vietnam should not wait passively. Instead, it must act proactively to build a legal system that is modern, responsible, and trustworthy.
Acknowledgement: While writing this article, the author used ChatGPT and QuillBot to check and revise grammar and sentence structure during the English drafting process.
By Nguyen Van Duong