AI can be incredibly useful, but if we are not careful, it can take us far beyond what is actually needed.
In Hancox v Sutherland & Ors [2026] EAT 139, the Employment Appeal Tribunal criticised a 300-page skeleton argument, said to be created using ChatGPT, as entirely unacceptable and said it served to obscure rather than illuminate. Reports on the case say it ran to almost 132,000 words.
What struck me is that this is not just a warning for litigants.
The wider point: This is a warning for all of us. AI is useful, but more detail, more angles and more words do not automatically produce a better result.
More is not always better
Whether it is a grievance, a disciplinary response, a tribunal document or a workplace investigation, the goal is not to cover every possible angle just because AI can.
The goal is to stay focused on what is relevant, accurate and proportionate.
And that is where the problem starts. AI can produce a huge amount of material very quickly, but volume is not the same as value.
The same risk applies to workplace investigations
Used well, AI can support structure, organisation and even evidence review.
Used badly, it can flood the process with points that do not really matter, distract from the actual issues and make decision making harder rather than easier.
Often that is not because the tool has failed. It is because we have not been clear enough about what we are asking it to do, and just as importantly, what we are not asking it to do.
For illustration only: useful boundaries might include:
-
stick to the issues actually raised and nothing else;
-
avoid speculation;
-
ask three specific questions which would answer the issue;
-
do not go beyond three paragraphs;
-
do not turn a reasonable investigation into an endless one; and
-
focus on what is necessary, not everything that is theoretically possible.
Clarity matters
That balance matters.
I do think there is too much easy criticism of other people’s use of AI without enough reflection on our own. Most normal users are still learning. I certainly think this applies across the board, including in HR, legal work and investigations.
I recently had 20 questions from a company client on a simple issue. I am 99% sure the client did not write those questions.
Final takeaway: The lesson from this case is not “don’t use AI”. It is use it properly. Slow down. Sense check it. Keep it relevant. Keep it proportionate.
If you would like to discuss the use of AI in HR processes, workplace investigations or employee relations, please contact Impact Lawyers.
By Victoria Hall, Co-Founder Employment Law
Victoria is an experienced employment lawyer, a Level 7 CIPD-qualified HR professional, accredited external workplace investigator, practising coach and a non-executive director.