The legal pitfalls of using Artificial Intelligence at the workplace

Businesses increasingly use artificial intelligence (AI) to speed up decision-making and HR processes, including recruitment, allocating work, managing performance and even dismissals. This guide for employers looks at how AI is used at work and the main legal risks to manage.

How AI is used in workplace decisions

• Profiling: algorithms analyse data to find patterns and make predictions about individuals, for example flagging employees who appear not to be meeting targets. This can lead to capability processes or dismissal.
• Automated decision-making: decisions are made by technology without human involvement. For example, a job applicant might complete an online personality test and be rejected automatically based on the result.
• Machine learning: systems learn from data to imitate human judgement. In recruitment, image or speech analysis may be used to assess candidates in video interviews.

These tools can save time, but they do not remove an employer’s legal responsibilities.

The risk of discrimination

AI systems are built and trained by people and on historical data, so they can reproduce existing bias. This is sometimes called algorithmic discrimination. The Equality Act 2010 prohibits discrimination because of a protected characteristic, and it applies to decisions made with the help of AI in the same way as decisions made by managers.

Employers should check what an AI tool is doing, review the data behind its outputs, and correct problems such as bias or important criteria being ignored.

Examples from practice

An Uber Eats courier brought an employment tribunal claim alleging that the facial recognition checks used to verify his identity were racially discriminatory, after his account was deactivated. The claim was supported by the Equality and Human Rights Commission and settled in 2024, with Uber Eats making a payment to the courier. The case shows how bias in AI tools can lead to legal claims.

Amazon reportedly abandoned an experimental AI recruitment tool after finding that it favoured male applicants. The system had learned from the CVs of people hired in the past, most of whom were men because of the male-dominated nature of the technology sector.

Indirect discrimination and accountability

In a situation like the Amazon example, a female candidate could bring a claim for indirect discrimination, because the tool puts women at a particular disadvantage. To defend the claim, the employer would need to show that using the tool was a proportionate means of achieving a legitimate aim.

Making recruitment more efficient may be a legitimate aim, but it is hard to see how a tool with such significant effects on candidates could be proportionate without meaningful human oversight.

Employers sometimes ask who is responsible when software introduces discrimination into a decision. Under the Equality Act 2010, the employer remains responsible for its recruitment and employment decisions, and the fact that the tool was bought from a third party is unlikely to be a defence.

Employers should also think about reasonable adjustments, as automated assessments can disadvantage disabled candidates. Our employment law advice for employers can help you assess these risks before you introduce a new tool.

Data protection and automated decisions

There is no UK law that bans the use of AI at work, but data protection law gives people some protection from automated decisions. Article 22 of the UK GDPR used to restrict decisions based solely on automated processing that have legal or similarly significant effects.

The Data (Use and Access) Act 2025 replaced Article 22 with a more permissive framework from 5 February 2026. Employers can now make significant solely automated decisions in wider circumstances, but must put safeguards in place, including:

• telling individuals about significant decisions made about them
• allowing them to make representations and challenge the decision
• allowing them to ask for human involvement in the decision

Stricter rules still apply where special category data, such as health information, is involved. Employers should also carry out a data protection impact assessment before using AI to make decisions about staff or candidates.

Regulation of AI in the UK and EU

The EU has adopted the AI Act, which treats AI systems used in recruitment and in managing workers as high risk and subjects them to specific requirements. Following amendments agreed in 2026, those high-risk obligations are due to apply from 2 December 2027. The AI Act can affect UK businesses that use AI systems in the EU or whose systems produce outputs used there.

At the time of writing, there is no UK legislation dealing specifically with AI in employment. Existing laws on discrimination, data protection and fair dismissal continue to apply. Trade unions have pressed for more protection: a 2021 report commissioned by the TUC recommended making it easier for workers to challenge discriminatory AI decisions and called for statutory guidance on safeguards.

Practical steps for employers

• Be open with staff and candidates about when and how AI is used.
• Keep meaningful human involvement in decisions about recruitment, performance management and dismissal.
• Test tools for bias before and after you introduce them, and ask suppliers how their systems have been tested.
• If AI informs scoring in a redundancy exercise, make sure the redundancy selection criteria remain objective, fair and explainable.

If you are introducing AI into your HR processes and want to reduce the legal risk, book a free 30-minute consultation or call our employment solicitors on 0141 471 9166.