A judge has said an Insolvency and Companies Court ruling should serve as ‘public admonishment’ of international firm Pinsent Masons and senior lawyers, after two botched uses of artificial intelligence.
ICC Judge Mullen concluded in Cork & Anor v Smith that the firm had misled the court twice: once by providing references which featured AI hallucinations, then a second time by using AI to produce an explanatory letter that was still wrong.
The judge said an unnamed junior lawyer (referred to as LA) seemed to have ‘almost entirely outsourced the thinking process’ to an AI program, while solicitor Samantha Poulton and partner Steven Cottee had failed to supervise LA properly.
Pinsent Masons has apologised and referred itself to the Solicitors Regulation Authority, which will also investigate whether any of the lawyers involved have breached the code of conduct.
The judge said it would be disproportionate to consider contempt proceedings but added that there was, at the very least, a prima facie case of a breach of the duty not to mislead the court and the duty not to waste court time.
He said: ‘Ms Poulton, Mr Cottee and, I will assume, LA, will have found this incident very embarrassing. Ms Poulton and Mr Cottee are experienced solicitors undertaking technical work in a reputable firm, no doubt under considerable pressure at times. I have no doubt that LA was working under pressure too. None of that excuses a failure to check the accuracy of the material that was placed before the court.’
Read more
The court heard that Pinsent Masons had been instructed to prepare a block transfer application for an insolvency matter. This would not normally require a hearing, but the judge gave directions for the matter to be listed following concerns about misleading statements made by the solicitors in two letters from March and April this year.
After the judge had queried an item with the firm, he was ‘taken aback’ by the response, which referenced a rule about the court’s power that he was not aware of and could not be verified.
He immediately assessed that the response had included an AI hallucination and wrote to demand an explanation.
The judge was then ‘astonished’ by the follow-up reply, which was again not credible and misapplied the law. The judge said that he had been concerned from the first response that a ‘cavalier attitude’ was being taken by Pinsent Masons to the accuracy of material put before the court, and the second response ‘only heightened my concerns’.
Instead of providing an immediate, full and truthful response to the first misleading letter, the judge suggested, the follow-up had ‘muddied the waters’ and was itself misleading.
Poulton and LA had day-to-day conduct of the matter. Poulton had been aware that LA used AI from time to time but did not know it had been used in this application. She accepted a failure on her part to supervise LA and check drafts properly, and is now doing a course to improve her understanding of AI.
Cottee, a partner for 14 years, was not involved in drafting either response and was unaware that AI was being used, but he was responsible for supervising Poulton and LA.
The court heard that LA started chatting with the firm’s AI following the judge’s query. Judge Mullen read these discussions and noted it was remarkable how, on a large number of occasions, the AI was ‘plainly wrong or, at the very least, extremely misleading’, with hallucinations beginning almost immediately. LA appeared not to have checked whether the AI responses were correct and asked the AI to draft a letter in response to the court.
The AI even tried to warn the junior solicitor that it was not ‘fully confident’ about the way it was interpreting the law, but this did not prompt them to check the text being quoted.
Both Cottee and Poulson checked the first draft letter going out to the court which contained the mistakes and had confirmed within around two hours that it was fine.
On the explanatory follow-up, LA again asked the AI to draft a letter for the court. The AI had even included an ‘unreserved apology’ in its draft, which it then took out after LA said the firm should not apologise. It was only after Poulton said there should be an apology that the LA copied her observation into the AI verbatim.
The judge said all three lawyers should have been aware of the dangers of using AI to conduct legal research. He was satisfied there was no intention to mislead on the part of Poulton and Cottee, and that LA had not appeared to have been candid with their supervisors about how the text was created. The junior solicitor’s conduct was found to be ‘very troubling’ and their apparent lack of care seemed ‘extraordinary’.
‘It was, however, incumbent on the more senior lawyers to check LA’s work,’ he added. ‘It also seems to me that, as part of the supervision of LA more generally, it should have been made clear to them, if it was not made clear, that they should check all references given by AI and be candid, when reporting the result of their research to their supervisors, as to whether AI had been used in that research and whether the content generated by the AI had been checked.’
He concluded: ‘AI has the potential to be wholly unreliable. AI may of course provide a jumping off point for research and legal reasoning but it does not, at least at present, do away with the need for proper research and thought on the part of a legal professional, even a very junior legal professional.’
Pinsent Masons has apologised and taken steps to put safeguards on its use of an internal AI pilot. The firm has also met its former client’s costs, including the costs of finding new lawyers.
