AI generated claims - good for BBQs perhaps, but what else?

The issue of how to handle AI generated claims and responses is hugely topical in both HR and claims management generally. Readers may have noted recent reports from Australia where a tribunal banned a claimant from using AI generated material, and expressed sympathy for the employer who had, as a result, stopped engaging with the issues. Could this spell the end for the use of AI in employment claims? Peter Woodhouse, Partner, and James Barratt, Associate Solicitor, discuss below the recent case from Australia, what Employment and HR teams are encountering, and what can be done to combat AI generated claims.

James reports on the case

“The Fair Work Commission is Australia's workplace relations tribunal. In a recent case, it ordered an employee to stop using AI to draft correspondence because it found that her AI-generated complaints had become ‘lengthy, wide-ranging, replete with generalisations, repetitive, and often couched in accusatory language’. This made it difficult for the employer to deal with her complaint, which had some genuine merit. The sheer volume of communication and its impenetrable nature put a stop to any meaningful conversation that could potentially have resolved matters and led to the employer effectively disengaging. The Commission therefore took the step of directing all future communications between the parties to be brief, accurate and respectful, and specifically prohibited the use of AI-generated correspondence. 
This is of course an Australian case. However, anecdotally, some Employment Tribunal judges are cautioning litigants in person of the limitations and downsides of using AI to present and argue their case. This is analogous to the grievance scenario, especially as non-addressed grievances often turn into Employment Tribunal claims.”
 

Peter investigates other experiences

“I asked our Employment and HR teams about their experiences of tribunal approaches. None of this is intended to describe a policy approach by UK tribunals, but rather to share practical experiences. This is what they said:”

• Respondents facing AI claims have a greater responsibility to take pro-active steps to identify issues.
• Litigants in Person, especially those who may have difficulty accessing or using non-AI engagement methods, quite rightly, attract sympathetic treatment from tribunal judges.
• Using AI to respond to an AI generated claim might not help because that can further obscure the real issues.
• One judge explained to a claimant that “AI may help you choose wine for a BBQ, but it's less useful for making a claim to the tribunal”.
• Sometimes the Respondent can produce a Scott Schedule to encourage the claimant precisely to identify the issues/claims.
• The most common and probably most helpful response is that in most cases, claimants are expected to understand and explain their complaints and often cannot do that if the claim is AI generated. Our employment judges are not yet as robust as the Australian judge in the case we mentioned, but we are detecting increasing sympathy for employers who require a complainant to do exactly that. 

What can you do?

I am sure many of us would sympathise with the employer who, when faced with a ‘war and peace’ grievance that leaves it none the wiser as to what the employee is concerned about, chooses to disengage. Employers should not do that as this can create employment risk, namely a constructive unfair dismissal claim (assuming they have qualifying service) and, depending on the facts, maybe a discrimination, victimisation, or whistleblowing complaint.  

Employers need to remember that AI can be very helpful. As an example, for employees with physical or mental health difficulties AI can provide a voice to the genuinely aggrieved. A blanket ban on the use of AI is unlikely to be appropriate. So what can an employer do? In our view, when presented with a discursive and excessively lengthy grievance, options to think about include:

• asking an employee to submit a succinct version, perhaps restricted to several pages (after all, Employment Tribunals often order witness statements to be restricted to a certain number of words);
• requiring a meeting to discuss the grievance before you can commence dealing with any matters; and
• asking directly about whether it involved AI and highlighting the issue with it, emphasising that it might not help the employee resolve their issue.

Effectively, the aim is to get to the heart of the AI generated grievance as quickly as possible, then once the issues are clearly identified, the usual grievance process can proceed as usual. It might be worth thinking about your grievance policy and whether it needs updating to cover AI generated grievances and associated correspondence, perhaps with restrictive template formats or limits on grievances. 

Our Employment Solicitors and HR Consultants are well versed in advising on employee complaints, with practical experience of how to deal with AI generated or supported complaints. If you'd like any advice or guidance about such a complaint that may be in your in-tray, please get in touch with a member of our team today. We also have an Employment Rights Hub, which has lots of useful information and resources related to recent past and upcoming changes to employment law.