Two more immigration lawyers facing potential disciplinary proceedings for misuse of AI

Two more immigration lawyers, a barrister and a solicitor, are facing the possibility of a Hamid hearing and a referral to their professional regulator after the Upper Tribunal issued show cause orders to each of them in cases involving alleged misuse of AI.

These are two unreported decisions, one was picked up by the media recently (while I was away) but it seems the second one wasn’t. I spotted it in this incredibly useful database of AI hallucination cases. The fact that such a database exists should give any decent lawyer serious pause for thought before using these tools, and to proceed with a great amount of caution where they do decide that the benefits outweigh the risks.

This article is also worth a read, not least because I suspect that many lawyers would be surprised to see some of the places where issues have arisen with legal research and AI.

Mr Chowdhury Rahman

The Upper Tribunal found it “overwhelmingly likely” that Mr Rahman used generative AI to draft the grounds of appeal. The grounds were riddled with errors and the tribunal also concluded that Mr Rahman had attempted to hide the fact of his AI use during the hearing. The case is ANPV and SAPV v Secretary of State for the Home Department UI-2025-003373 & UI-2025-003374.

The appellants are Honduran sisters aged 25 and 19 who were appealing the First-tier Tribunal’s decision to reject their appeal against the refusal of their asylum claim. They instructed their barrister Mr Chowdhury Rahman on a direct access basis.

The Upper Tribunal found that there was no error of law in the First-tier Tribunal’s decision and the appeal was dismissed. Unfortunately for Mr Rahman, there was a postscript to the decision headed “Inaccuracies in the Grounds of Appeal”.

The judge set the appeal grounds out in full in decision and said that some of the assertions made about the First-tier Tribunal’s decision were “simply incorrect”. This was not the only problem. The judge went on to say that:

33. Twelve authorities are cited in the grounds. When I came to read the papers in this appeal, I was concerned that some of those authorities did not exist and that others did not support the propositions of law for which they were cited in the grounds. When the appeal was called on before me yesterday, therefore, I directed that Mr Rahman should return to his chambers in Lincoln’s Inn and provide me with printed copies of the authorities, with the relevant paragraphs of those authorities side-lined. Given that Mr Rahman had settled the grounds in July 2025, and given that he was instructed to appear before me to make submissions on his own grounds, he should have been in a position to take me through the authorities immediately, but I gave him four hours in which to comply with the direction.

34. Mr Rahman returned with a number of authorities at 2pm, and I asked him to take me to the authorities in order, so that he could identify the passages which were said to support the legal propositions in the grounds. He was unable to find some of the authorities. Of the authorities which he was able to find, there was not one which offered any support for the propositions of law which were set out in the grounds. On a number of occasions, Mr Rahman said that he had made a mistake and that he had intended, instead of citing one authority, to cite a completely different one. Often, however, the authority he said that he had intended to cite was also irrelevant to the proposition of law set out in the grounds.

35. There were a number of occasions on which Mr Rahman was unable to locate on Bailii electronic copies of the authorities which he had provided in hard copy. Throughout, it seemed that he was unfamiliar with Bailii or any other legal search engine. Mr Rahman was consistently unable to grasp the point that it was the ratio decidendi of the case to which I should be taken; oftentimes he took me to parts of the argument, or even to the facts of the case. At the risk of stating the obvious, counsel who comes to argue any case should be able to identify the salient paragraphs of the authorities cited in the grounds of appeal, but that is all the more so when the grounds of appeal were recently drafted by the same member of the Bar. Mr Rahman was however so unfamiliar with the cases, even after the additional four hours he was given, that the process of going through the authorities in an attempt to locate the relevant passages took more than two hours.

The tribunal then went through the first ten authorities cited, along with what the barrister said about it after returning from the break. Some of the cases did not exist, others that did exist were not relevant to the grounds of appeal. The tribunal said that “All of the submissions which were made in the grounds were therefore misleading” and concluded that:

The most obvious explanation is the first possibility: that the grounds of appeal were drafted in whole or in part by generative Artificial Intelligence such as ChatpGPT. I am bound to observe that one of the cases cited in Mr Rahman’s grounds – R (YH) v SSHD [2010] EWCA Civ 116 – has recently been wrongly deployed by ChatGPT in support of similar arguments concerning section 8 of the Treatment of Claimants Act: see MS (Professional conduct; AI generated documents) Bangladesh [2025] UKUT 305 (IAC).

Mr Rahman had said that he used Microsoft Copilot to summarise decisions but he did not produce any of these summaries to the tribunal. The tribunal used Copilot to replicate a summary of one of the cases and found the summary to be accurate, stating that while it was “not on a par with the headnote to an official law report, it is at least faithful to the subject matter of the case”.

As far as next steps are concerned, the tribunal said that “I am minded, in light of the misleading statements in the grounds and the resulting waste of the tribunal’s time, to refer Mr Rahman’s conduct to the Bar Standards Board. I will make a Show Cause order, requiring him to set out in writing why that course should not be followed”. It is possible that a Hamid hearing will then be listed.

Mr Rahat Ullah

The second case involved a solicitor, Mr Ullah of Morgan Lane Solicitors and the case is AK v Secretary of State for the Home Department UI-2025-002918. The appeal itself was against the refusal of asylum and human rights claims.

The First-tier Tribunal dismissed the appeal and the appellant’s solicitors submitted grounds of appeal to the Upper Tribunal with Mr Rahat Ullah’s name on them. When it came to preparation of the hearing bundle, amended grounds of appeal were included instead of the original grounds, and part of the third ground of appeal had been removed. The index described this document as “Grounds on which the appellant has been granted permission to appeal”.

The Upper Tribunal dismissed the appeal, finding that there was no error of law in the First-tier’s decision. Again, there was an ominous post-script headed “Show cause notice”.

The tribunal had directed Mr Ullah to explain why he had amended the original grounds of appeal. Mr Ullah responded to that direction accepting that he had amended the grounds and that the index was inaccurate. He apologised to the tribunal, which said:

3. Mr Ullah’s response misses the point which was apparent to Mr Raza at the hearing. The difficulty with the grounds is not merely that sentences were removed; it is why those parts of the original grounds were removed. The cases cited in the original grounds – Nye v SSHD [2021] EWCA Civ 148 and R (Goudey) v SSHD [2012] EWHC 567 (Admin) – do not exist. As I observed to Mr Raza at the hearing, I have however come across both of those citations in the past, in cases in which practitioners have failed to check the accuracy of drafting undertaken by generative Artificial Intelligence such as ChatGPT.

4. A failure to check the accuracy of such drafting can result in a court or tribunal being misled and is potentially grounds for referral to the Solicitors Regulation Authority or even stronger sanction: R (Ayinde) v London Borough of Haringey and Al-Haroun v Qatar National Bank and Anor [2025] EWHC 1383 (Admin).

5. What concerns me in this case is not merely that there were false citations in the grounds of appeal considered by Judge Saffer; it is that those false citations were then removed from the grounds of appeal which were placed in the composite bundle. The former actions are unprofessional, the latter are potentially dishonest because it suggests that there was an attempt to conceal the false citations. It was for those reasons that I made it clear to Mr Raza that I expected there to be a full response from Mr Ullah.

Mr Ullah was directed by the tribunal to show cause as to why his conduct in this matter should not be referred to the Solicitors Regulation Authority.

Conclusion

I think it is worth noting that both cases here were heard by the same Upper Tribunal judge as in the other recent case. One hopes that other judges are equally alive to spotting these things and making Hamid referrals, as this sort of behaviour needs to be stamped out as efficiently and quickly as possible, although it does seem rather unlikely that all of the misuse of AI cases have ended up in front of the same judge.

Many people describe the use of AI as “inevitable”, but I do not agree and I certainly have no intention of using it, least of all in my legal writing or in any work for my clients. I personally don’t think that there are many (if any) shortcuts to doing a good job as a lawyer and I do not believe that AI is a solution to time and resource pressures, not least because of the number of mistakes it makes, which mean that any work generated in this matter has to be thoroughly and carefully checked – a time consuming exercise which undermines the reason for using it to begin with.

There is obviously also the risk of de-skilling yourself, as well as a large number of ethical concerns (including but not limited to environmental concerns, copyright theft and biases). However others disagree, which is why we will continue to see these decisions.

Last, but certainly not least, is the people at the heart of these cases. Both cases here involved protection claims, so if meritorious, lives were on the line. What I fail to see is how any of this benefits the clients. Perhaps this is an aspect that will be further examined by the regulators, should things get to that point – but some of the questions I have is what were clients told about the use of AI in their cases? Were they offered a choice? Assuming the reason for using AI was to reduce the amount of time spent on the work – were their fees actually reduced? Is there evidence of any of this on file? I hope that the regulators do act on these or similar cases with speed.