The rise of AI-generated workplace disputes: What HR needs to know
AI is transforming the way employees raise workplace complaints. Over the past 12 months, employers have seen a sharp rise in AI-assisted grievances and Employment Tribunal claims. While AI can help employees articulate genuine concerns, it can also generate lengthy, aggressive and factually inaccurate submissions that make resolution harder. Below, we set out the key challenges employers are facing and practical steps HR teams can take in response.
AI-assisted grievances
AI can help employees summarise their complaints, but in practice, AI-assisted grievances often make the investigation process harder. Here are the most common issues employers are encountering - and potential solutions.
AI can generate multiple pages in seconds. Some employers are seeing grievances running to hundreds of pages as a result.
In response, employers are introducing suggested word or page limits on grievances. This is not without precedent - some Employment Tribunals already impose page limits on document bundles or word limits on witness statements. Exceptions may sometimes be necessary, but a suggested limit can be an effective way to keep grievances manageable.
Others are adopting a standard form on which grievances must be submitted. Again, exceptions may need to be made for particularly complex matters.
Some employers already provide that they will not consider grievances about very historic events. To bring a Tribunal claim, ACAS must be notified within three months (soon to be six months) of the event complained of. In certain claims, such as discrimination, the Tribunal can extend time where there has been a continuing series of events or where it is just and equitable to do so. Employers may therefore need to allow for some exceptions, but many are considering implementing a general expectation that grievances should relate to recent events.
It is not uncommon for AI-generated grievances to contain factual errors where the employee has not properly reviewed the output. In response, employers are adding to their grievance policies a reminder that employees must take accountability for the content of any grievance they submit. For example, if a grievance makes a serious and factually incorrect allegation about a colleague, it will not be sufficient for the employee to say, "That was AI, not me.”
A common criticism of AI-generated content is that it is overly confident and assertive, particularly when setting out legal claims. The tone can lack nuance, creating an adversarial atmosphere from the outset and reducing the chances of an amicable resolution. The ACAS Code of Practice on Grievances requires employers to arrange a meeting with the employee after receipt of the grievance. This meeting becomes even more important in the context of an AI-assisted grievance: meeting with the employee and asking them to explain their complaint provides an opportunity to identify the employee’s real concerns and explore solutions before positions become polarised.
We recommend that employers add to their Grievance Policy (and also have a Generative AI policy) clarification that employees should not, for any purposes, upload employer confidential information or personal data to public AI platforms. Employers may warn employees that doing so could lead to disciplinary action.
This does not necessarily prevent employees using AI to assist them as they can use appropriate privacy settings or anonymise any information that they upload, but to protect the personal data of colleagues, employers should be clear to employees about this issue.
Compliance with data subject access requests is important, but it is worth seeking early advice in relation to lengthy requests that have been AI-assisted as it may be that the scope of requests can be narrowed.
Get in touch
AI-assisted Employment Tribunal claims
We are aware that some legal businesses are using AI to varying degrees in bringing Employment Tribunal claims. Many litigants in person are also using it and judges have indicated that, when used properly, AI can be a powerful tool to help claimants assess and structure their claims. However, there are currently many examples where its use has been far from constructive.
As with grievances, AI makes it easy to generate a very long claim and employers may then feel obliged to respond to every detail.
Whilst it can be difficult for employers and their representatives facing a lengthy claim form, they may want to consider the comments of the EAT in C v D (2019): “I do not encourage parties, particularly lawyers, to engage in that type of 'narrative' pleading. I would encourage legal representatives, in particular, to adopt a more succinct and clear drafting style…..A claim form sets out a legal claim. It is not a witness statement…..the narrative style of pleading makes the task of Employment Judges who need to case manage the case more difficult: it takes more time than may be available to properly identify the issues”
A long claim form may look impressive, but often the substance of the actual claims can be weak. Employers should concentrate on the legal claims and not irrelevant detail. Employers can also use the Tribunal’s case management powers to their advantage: for example, by seeking a strike-out or deposit order in respect of weak claims.
A growing issue in practice is AI-assisted claimants making multiple and frequent applications to the Tribunal - taking every possible opportunity without discernment as to what is appropriate. Claimants in such cases should be reminded of the Tribunal’s overriding objective, which includes dealing with cases in ways proportionate to the complexity and importance of the issues. Those making frequent and disproportionate applications (whether claimants or respondents) risk subsequent applications for the recovery of legal costs wasted in responding.
The Employment Tribunal has recently issued Presidential Guidance on Interim relief applications, noting a marked increase in applications seeking this rare form of emergency relief from dismissal, “often indicating use of artificial intelligence”. The volume of such applications that Tribunals would previously have received in a year, they are now receiving every month, accompanied by a significant increase in supporting documentation. The guidance does not prevent applications being made, but does make claimants aware of the high hurdles they must clear to obtain this emergency relief in the hope that the volume will decrease.
Where AI is reliant on information provided by a Claimant, or is incorrectly prompted, it can provide a one-sided view of the merits of a claim, together with an unrealistic assessment of what compensation might be awarded. This issue is likely to be magnified when the cap on unfair dismissal compensation is removed in January 2027. Employers seeking to settle claims on a pragmatic and commercial basis can find it harder to do so where claimants have unreasonable expectations of the likely outcomes. Employers should consider the use of ACAS, Judicial Mediation or Judicial Assessment to seek to manage claimant expectations and provide a landscape in which settlement is possible.
Privilege
Legal privilege is a concept that permits parties to seek legal advice without that advice becoming public. An employer can ask its lawyer to report on the merits of its defence without concern that the Claimant will be able to seek disclosure of that advice.
There is considerable debate about whether "privilege" is available to parties using AI to assist with their claim or defence. A key requirement of privileged information is that it must be confidential. If parties are seeking advice from public AI platforms, or uploading documents to AI, that information is arguably no longer confidential. To the extent such information is relevant to the claim or its defence, an employer may be able to seek disclosure of an employee’s AI prompts and the responses received.
The use of AI in workplace disputes will only increase. The employers best placed to manage these challenges are those taking proactive steps now: reviewing grievance policies, setting clear expectations around AI use and data confidentiality and ensuring that managers and HR teams are equipped to look beyond AI-generated content to identify and resolve the real issues at stake.
How we can help
If you have any queries regarding the content of this legal article, please get in touch with authors Rena Magdani, Matt McBride or another member of our Employment team.
Read the other articles featured in our latest Employment update here:
The content of this page is a summary of the law in force at the date of publication and is not exhaustive, nor does it contain definitive advice. Specialist legal advice should be sought in relation to any queries that may arise.
Related news & articles
Related expertise
Contact us today
Whatever your legal needs, our wide ranging expertise is here to support you and your business, so let’s start your legal journey today and get you in touch with the right lawyer to get you started.
Get in touch
For general enquiries, please complete this form and we will direct your message to the most appropriate person.