Trade Marks
What happens when ChatGPT meets trade mark proceedings: a recent misstep at the UKIPO
July 03, 2025
A recent appeal of a UK Intellectual Property Office (UKIPO) decision resulted in both parties receiving disparagement after submitting skeleton arguments that were at least partially generated using ChatGPT. The full decision can be found here - but what happened to the parties?
When ChatGPT was used - and where it went wrong
ChatGPT was used by both a litigant in person and a legal representative when undertaking legal research and when drafting skeleton arguments.
The litigant in person’s skeleton arguments contained both fictitious citations and real - but unsupported - citations. Meanwhile the legal representative’s skeleton arguments had issues with legal authorities. In short, ChatGPT appears to have ‘created’ case law that does not exist and attributed quotes to cases that do not contain them.
Risks of using ChatGPT
It is therefore clear that there are explicit risks when using artificial intelligence (AI) - such as ChatGPT - for legal research and the drafting of skeleton arguments, as it does not always provide accurate information.
Using AI to create a statement of case can lead you to rely on entirely fictious arguments, as ChatGPT - for example - is not capable of conducting reliable legal research. As noted by the Divisional Court, although responses generated by ChatGPT may appear “apparently coherent and plausible”, these “coherent and plausible responses may turn out to be entirely incorrect”.
While the foundation of your argument may remain valid – if, by chance, your points are supported by real case law - the ‘punch’ of your argument is significantly dampened. Furthermore, if you rely on AI when drafting skeleton arguments, you are less likely to be familiar with your skeleton arguments. Therefore, if your skeleton arguments become closely scrutinised (due to them relying upon AI), the UKIPO may be less likely to follow your case - after all skeleton arguments are meant to help guide the Appointed Person towards the ‘correct’ conclusion.
So, what happened?
As a result of both parties using artificial intelligence, the UKIPO was forced to consider, inter alia, referral to the police, contempt, strike out and wasted costs.
In light of all the circumstances, the litigant in person’s actions did not result in any sanctions being imposed. However, if such conduct were to be repeated in the future by a litigant in person, they are unlikely to receive such a lenient outcome. The representative’s actions were also treated generously, and they were not referred to IPREG (the Intellectual Property Regulation Board).
While both parties were let off lightly (in my view), this decision clearly outlines the range of sanctions available in such scenarios - and now that they have been set out, harsher consequences are likely to follow in future instances.
Points to take away
Whether or not you are experienced with trade mark proceedings, it is always advisable to seek professional legal guidance from a trade mark attorney and to ensure that your representative conducts accurate legal research. If you are about to begin proceedings at the UKIPO, or are amid them right now, do not hesitate to seek our support.