Running a business is hard enough without the added pressure of navigating an employment law landscape that feels like it shifts every few months. But right now, there’s a new layer of complexity that we’re seeing trip up even well-intentioned employers: the rise of AI tools like ChatGPT and Microsoft Copilot being used, by both employees and employers, to get HR and legal guidance.
The problem isn’t that AI exists. The problem is that it sounds convincing, even when it’s wrong.
Here are five HR situations we’re seeing businesses get wrong right now, and what to do instead.
1. Letting AI answer your employees’ HR questions, before you do
We recently worked with a business whose employee came into a meeting armed with a printout from ChatGPT telling her she had grounds for an age discrimination claim. She didn’t, but the business had no idea how to respond, and the situation quickly escalated into something costly and stressful.
AI tools are brilliant at making things sound authoritative. They’re not brilliant at understanding the nuances of your specific situation, your employment contracts, or the current state of UK employment law. When employees start relying on ChatGPT for HR advice, the risk lands squarely with you.
What to do instead: Make sure your employees know who to come to with HR concerns. A clear, accessible process, and an open-door culture, reduces the chances of an employee turning to Google (or ChatGPT) before they come to you.
2. Underestimating how digital communication creates risk
Would an aubergine emoji sent in a work WhatsApp group land your business in an employment tribunal? Possibly, yes. The line between casual banter and harassment has never been thinner, particularly in digital channels that blur the personal and professional.
Businesses are regularly caught out by conduct in messaging apps, Teams channels, and group chats that they didn’t know was happening, and had no policy to address. By the time it becomes a grievance, the damage is already done.
What to do instead: Review your communications and social media policy. Make sure it covers all digital channels, not just company email. And train your managers, they’re often the ones who witness this behaviour first.
3. Assuming protection only applies to diagnosed conditions
One of the most common, and costly, misconceptions we encounter is the belief that an employee needs a formal diagnosis before they’re protected under the Equality Act. They don’t.
If an employee has a suspected or perceived neurodivergent condition, ADHD, autism, dyslexia, and you treat them less favourably because of it, you may already be exposed. The law can apply even where no diagnosis exists, particularly if you were aware (or should have been aware) that a condition might be present.
What to do instead: Approach performance and conduct issues with curiosity before you reach for the disciplinary process. If something feels off, ask questions, offer support, and consider whether a referral to occupational health is appropriate.
4. Using AI to manage redundancy processes
We’ve seen a sharp increase in the number of employees using AI to support them through redundancy processes, drafting consultation questions, writing appeals, and preparing for meetings. On the face of it, that seems reasonable. But it’s creating real complications for employers.
AI-generated consultation responses tend to be longer, more legalistic, and sometimes more aggressive in tone than employees would naturally produce themselves. They can create an overwhelming volume of questions, many of which are irrelevant, and risk derailing a process that should be straightforward. Meanwhile, employers making redundancies are sometimes turning to Copilot or ChatGPT for guidance on the process itself, and receiving advice that is out of date, jurisdiction-blind, or simply incorrect.
What to do instead: Get proper HR support before you start any redundancy process. The rules around consultation, selection criteria, and notice are specific, and getting them wrong is expensive. Read our full guide on navigating redundancy risk in the age of AI here.
5. Doing nothing because you don’t know what to say
This might be the most common mistake of all. When something difficult lands in a business, a grievance, a capability issue, a sensitive conversation about health or behaviour, many managers freeze. They don’t know the right words, worry about saying the wrong thing, and so they say nothing.
Silence isn’t safe. Delays in addressing issues often make them worse, give employees the impression that complaints aren’t being taken seriously, and can undermine your position if a case ever reaches a tribunal.
What to do instead: You don’t need a perfect script, you need a clear process and the confidence to follow it. That’s what good HR support gives you. If you’re not sure what to say, pick up the phone to us before the situation escalates.
The bottom line
AI isn’t going away, and neither is the complexity of managing people. But the two together are creating a perfect storm for businesses that don’t have proper HR support in place. The situations above aren’t rare edge cases. They’re the conversations we’re having with businesses every week.
If any of them sound familiar, let’s talk, speak to our team today or book a free HR consultation here.
Holly Mapstone is an HR expert and specialises in helping those of you in the SME sector with all of your HR needs.
Consultancy allows Holly to be agile and to adapt her style and support to suit the needs of her clients. Building strong working relationships allows Holly to deliver improvements, while also aligning people strategy to business outcomes and promoting resilient and positive cultures.
