BLOG
Don’t rely on AI | 5 areas where employers should exercise caution
Written by Amy Waters on 20 June 2025
Artificial Intelligence (AI) tools are transforming the way employers manage their operations, from streamlining recruitment to drafting employment documents.
While the benefits of AI are undeniable, its use in the workplace must be carefully controlled. Relying too heavily on AI – especially without human oversight or clear policies – can lead to legal, ethical, and reputational risks.
Here are five employment-related tasks employers shouldn’t rely solely on AI for.
1. Shortlisting candidates
Many employers, especially larger organisations, use AI tools like CV-ranking software and keyword scanners to sift through applications quickly and efficiently. However, while AI can speed up the recruitment process, it should never replace human oversight.
A cautionary example is Amazon, which scrapped its AI tool after it was found to downgrade CVs from graduates of women-only colleges, introducing gender bias. This shows how AI, trained on historical data, can unintentionally reinforce discrimination.
Using AI in shortlisting isn’t unlawful, but employers must proceed carefully. Under the Equality Act 2010, individuals are protected from discrimination even at the pre-employment stage – so rejected applicants can bring claims if AI disproportionately disadvantages protected groups. The fact that a decision was made by an algorithm does not shield employers from liability; Employment Tribunals are likely to hold employers responsible, even without discriminatory intent.
To reduce risk, employers should sample rejected applications to check for bias or disproportionate outcomes. Such reviews are permitted under UK data protection law if designed to detect errors and must meet ICO standards. Employers must also disclose AI use in their Privacy Notice and document it in a Data Protection Impact Assessment to ensure transparency and compliance.
Do you need support?
Speak to us for an honest, no obligation chat on:
0345 226 8393 Lines are open 9am – 5pm
2. Drafting employment documents
AI can be a helpful tool for creating first drafts of employment contracts, policies, and procedures. However, it can sometimes mix up legal requirements from different countries or jurisdictions, which poses significant risks.
The wording of your contracts and policies is critical in employment disputes or claims – and you don’t want to discover they don’t hold up when it’s too late. For this reason, AI should only be used to create a first draft, with all documents thoroughly reviewed and finalised by a qualified legal professional to ensure accuracy and compliance.
This is particularly important when it comes to complex documents like settlement agreements, which must include specific legal phrasing in order to be valid. If an agreement isn’t compliant or worded correctly (due to being written by AI), it may be legally unenforceable – leaving the door open for the employee to still pursue claims.
3. Making dismissal decisions
Many employers now use AI tools to monitor productivity by tracking metrics like email output, calendar usage, and keyboard activity. While AI can flag potential concerns, relying solely on automated data to dismiss employees is extremely risky and can lead to unfair dismissal claims.
This risk extends to redundancy decisions. Using AI to select which employees are made redundant – or even partially basing decisions on AI-generated rankings – can raise serious legal issues, including claims of algorithmic bias and discrimination. Under UK GDPR, employees have the right to know if decisions affecting them are made by automated systems, to appeal those decisions, and to request human review.
Moreover, the Employment Rights Act requires dismissals to follow a fair and reasonable process. This includes giving employees a proper hearing where they can respond, and ensuring a manager – not an algorithm – makes the final decision. Acting solely on AI-generated alerts or recommendations without meaningful human involvement is unlikely to meet this standard and could expose employers to legal challenges.
4. Improving efficiency by replacing older workers
AI can drive efficiency by automating roles, but employers must be very careful when these changes disproportionately affect older employees. Take IBM’s example: the company publicly announced it had replaced hundreds of HR roles with generative AI agents. While the media praised this as innovation, two former HR employees – both over 60 – filed age discrimination claims, citing this announcement as evidence they were pushed out to make way for AI and younger hires.
This case is a cautionary tale. Even if automation is a legitimate business decision, UK law requires employers to avoid any actions or communications that could be seen as discriminatory toward protected groups. When efficiency improvements result in older workers being disproportionately replaced, Tribunals will closely examine whether the employer managed the process fairly and without bias.
Importantly, how leadership talks about AI-driven workforce changes matters just as much as how those changes are implemented. One careless comment or public statement can become critical evidence in a discrimination claim.
5. Providing employment advice
AI tools like ChatGPT offer quick, tailored responses to your specific questions, making them feel like a more personal, efficient alternative to trawling through Google. This speed and apparent relevance can make AI seem like an easy solution when employee issues arise.
However, AI isn’t a substitute for professional HR or legal advice. While its answers may sound confident and convincing, they can be inaccurate, outdated, or based on biased data. Acting on this flawed information can lead to poor decisions, legal missteps, and significant costs.
AI can assist – but it can’t replace the insight, judgement, and accountability that only qualified professionals provide. Always consult the experts, particularly when dealing with complex or sensitive issues – otherwise, what seems like a money-saving shortcut could end up being far more costly in the long run.
Related Content
Protect your workplace from AI-related risks with WorkNest
Whether you need help managing the fallout from AI or getting on the front foot, our Employment Law and HR experts are here to help.
We can support you in limiting the risks of AI use, managing employee issues like misuse or misconduct, and crafting clear, future-ready policies to prevent problems before they arise.
For advice and assistance, get in touch on 0345 226 8393 or request your free consultation using the button below.
