A barrister who submitted a fake legal case, generated by ChatGPT, during a tribunal has claimed he was a victim of the technology’s habit of hallucinations.
Muhammad Mujeebur Rahman, who was representing an appellant in an immigration appeal, cited a non-existent case, Y (China) [2010] EWCA Civ 116, claiming it supported his argument around procedural delay.
ChatGPT Use Uncovered
The issue came to light during a hearing in June, when the barrister was questioned by the tribunal about the case. He initially told the panel he had intended to refer to other, genuine cases, and claimed the incorrect citation was a mistake.
However, after being given time over the lunch break, the barrister returned to insist that the case was a legitimate Court of Appeal judgment, and went on to name the supposed panel of judges. He then submitted a nine-page document to the tribunal clerk that included some misleading statements, falsely linking the citation to a real case and failing to reference any relevant authority on delay.
The tribunal panel have the barrister a deadline to either provide a legitimate copy fo the judgment or explain his error. In a letter submitted before the deadline, he admitted he had meant to cite a different, real case, but had failed to provide the correct reference. He attributed the mistake to a combination of illness and a recent hospital stay.
AI Error and Accountability
At a subsequent hearing, the barrister admitted for the first time that he had used ChatGPT to draft the original grounds of the appeal and to generate the document handed to the tribunal. He argued that he had been misled by the AI tool and was “also a victim”, saying he believed the fabricated case to be genuine.
The panel rejected this line of defence, stating that the barrister had made no attempt to verify the case through reliable research platforms such as Westlaw, LexisNexis, BAILII or EIN. It concluded that his earlier explanations had been “a less than honest attempt to pretend the citation was a typographical error.”
While the tribunal did not find evidence of deliberate fraud, it said the barrister’s conduct fell short of professional standards and likely influenced the tribunal’s decision to grant permission on one of the grounds. It ruled that his actions breached core duties of honesty and integrity and confirmed that the matter would be referred to the BSB.
Tribunal Issues Warning on AI Use
The panel made clear that legal professionals have a duty to verify the accuracy of authorities cited in proceedings:
“Taking unprofessional short-cuts which will very likely mislead the Tribunal is never excusable,” it warned.
The barrister has since completed further training on immigration law and the responsible use of AI. He apologised for his actions and asked the panel not to escalate the matter to the regulator, arguing that he now understands his obligations and will act with integrity going forward.
Despite his apology, the tribunal upheld the referral, underlining the seriousness of misusing AI tools and submitting fictitious legal authorities in judicial proceedings.
