Even in a week when Arsenal clinched the Premier League and Thomas Tuchel announced the England World Cup squad, the Spygate saga is perhaps football’s biggest talking point.
Southampton were thrown out of this weekend’s play-off final against Hull City and deprived of the chance to win the £200m lottery ticket that is promotion to the Premier League. This was after the club admitted to spying on play-off semi-final opponents Middlesbrough as well as other clubs during the main season.
Southampton admitted wrongdoing but appealed the sanction, arguing it was disproportionate in comparison with the £200,000 fine handed to Leeds United after a similar offence in 2019.
The comparison is flawed. When Leeds spied on Derby County seven years ago, there was no specific rule preventing it (the club was instead charged with not treating another club with the ‘utmost good faith’). After the incident, the Football League brought in a regulation banning clubs from observing a rival’s training within 72 hours of a match.
In other words, Southampton knew they were breaking the rules. But crucially for the appeal, they didn’t know the penalty. The EFL did not express any certainty that a breach could lead to expulsion or a points deduction. For what it’s worth, I think the punishment is merited but also dangerous. All teams cheat, to a greater or lesser extent. The fact this was so obvious and easy to prove appears to have led to the harshest result. But the lack of a defined outcome to fit the crime was an oversight.
So what, I can hear my news editor plead with exasperation, does this have to do with the legal profession? Well, we’re (very slowly) getting there.
I think there are parallels here in how we should approach the regulation and enforcement of improper use of AI in litigation.
The professional bodies and regulators are belatedly issuing guidance and advice on what is acceptable use of generative AI in preparing court documents. This follows a spate – which shows no sign of abating – of cases where qualified lawyers have presented false or irrelevant authorities to the court in support of their argument. These appear to have been produced as a result of searches using AI to craft a submission.
This really should be a simple thing to define and to prevent. The use of AI to draft a document is acceptable, but the result must be something you are prepared to stake your professional reputation and regulatory status on. It is no different to asking a paralegal to draft a document: in theory this is fine (albeit the client might want to know for the purposes of fees), but the responsibility for checking it is correct remains with the person who is signing off on the document. To simply allow a document to go before a court without the proper vetting and oversight is lazy and negligent.
The regulators should step in now and make clear that presenting false authorities is unacceptable and – most importantly – will result in disciplinary action. Like those Southampton analysts, lawyers relying on AI to draft documents know what they are doing is reckless and lacking integrity. But the best way to stop it is for them to know it has consequences. It should come as no surprise to the next person doing it when they appear before the disciplinary tribunal.
The lesson from Spygate is to set the boundaries and the outcomes before people are tempted to cut corners. Lawyers do not need vague warnings and retrospective outrage: they should be told exactly where the line is and what will happen if they cross it.
1 Reader's comment