d&i Leaders Equality Law Update
AI bias in recruitment and promotion: navigating legal and discrimination risks
By Rupa Mooker on 20 June 2025
Artificial intelligence (AI) is rapidly transforming the way organisations in the UK recruit and promote staff. From automated CV screening to video interview analysis to promotion recommendations, AI tools promise efficiency, consistency, and the ability to process vast amounts of data quickly. However, these systems also carry significant risks, particularly the potential to replicate, or even amplify, existing biases. Understanding the legal and ethical implications of AI in employment decisions is crucial for those involved in them. So, what are the potential risks of AI bias in recruitment and promotion and is current UK employment law equipped to address these challenges?
The rise of AI in employment decisions
AI is increasingly used at various stages of the employment lifecycle. Algorithms scan CVs for keywords, qualifications, and experience, shortlisting candidates automatically. Those getting through to a video interview may then be subject to AI analysis of their facial expressions, tone of voice, and word choice to assess suitability. When looking forward to the next batch of promotions, businesses are increasingly using data-driven tools to evaluate employee performance and potential, suggesting who should be considered for advancement.
While these technologies can help reduce human error and speed up decision making, they are not immune to bias. In fact, if not carefully designed and monitored, AI systems can entrench, and even worsen, existing inequalities. A report published in November 2024 by the UK Information Commissioner’s Office (ICO) raised concerns over the use of AI powered recruitment software, revealing that candidate applications were being filtered based on protected characteristics such as gender, race, and sexual orientation. It highlighted that the very systems being used to try and prevent discriminatory practices by monitoring biases, can inadvertently infer candidates’ protected characteristics using the data contained in their applications. According to the ICO, this inferred information is often inaccurate, frequently processed without candidates’ knowledge or consent (raising concerns around data protection legislation) and does not effectively guard against bias.
Ian Hulme, ICO Director of Assurance said “AI can bring real benefits to the hiring process, but it also introduces new risks that may cause harm to jobseekers if it is not used lawfully and fairly.” He also said, “Organisations considering buying AI tools to help with their recruitment process must ask key data protection questions to providers and seek clear assurances of their compliance with the law.” The 6 key questions that the ICO advises businesses to ask before procuring an AI tool for recruitment can be found here.
Understanding AI Bias
AI bias occurs when an algorithm produces results that are systematically prejudiced due to flawed assumptions in the machine learning process. This can happen in several ways. For example, if the data used to train an AI system reflects historical inequalities such as fewer women, or people from ethnically diverse backgrounds in senior roles, AI may learn to favour certain groups over others. The way an algorithm is programmed can also inadvertently prioritise certain characteristics such as educational background which may, for example, disadvantage underrepresented groups. If these biased outcomes are then fed back into the system, the AI can become increasingly skewed over time. For instance, an AI system trained on CVs from a male dominated industry may learn to prefer male candidates, even if gender is not an explicit factor. Similarly, video interview tools may misinterpret accents or facial expressions, disadvantaging candidates from certain backgrounds.
Discrimination law and AI bias
The UK’s primary legislation on workplace discrimination is the Equality Act 2010 which protects individuals from discrimination based on nine protected characteristics – age, disability, gender reassignment, marriage and civil partnership, pregnancy and maternity, race, religion or belief, sex, and sexual orientation.
Under the Equality Act, discrimination can be direct (treating someone less favourably because of a protected characteristic) or indirect (applying a policy or practice that disadvantages people with a protected characteristic, unless it can be objectively justified).
The Equality Act does not specifically mention AI or automated decision-making. However, if an AI tool leads to discriminatory outcomes, employers can still be held liable. For example, if an AI system disproportionately rejects candidates from a particular ethnic group, this could amount to indirect discrimination unless the employer can show their system is a proportionate means of achieving a legitimate aim. The law also requires employers to make reasonable adjustments for disabled candidates and employees. If an AI tool disadvantages disabled people (for example, by penalising gaps in employment due to illness), this could breach the duty to make reasonable adjustments.
Whilst the Equality Act provides a foundation for challenging AI driven discrimination, there are gaps and uncertainties. The lack of transparency raises issues as AI systems are often set up in a way that makes it difficult to understand how decisions are made or to identify the source of bias. It can also be unclear who is responsible for biased outcomes – is it the employer, the software provider, or both? The law is currently largely reactive, relying on individuals to bring claims after discrimination has occurred, rather than requiring employers to proactively audit and mitigate AI bias.
Preparing for legal and regulatory change
Businesses should anticipate that legal requirements will become more stringent in the coming years and start to prepare for this. Here are some proactive steps to consider:
- Audit and monitor AI systems
Businesses should regularly review their AI tools for potential bias, both before implementation and throughout their use. Testing systems with diverse data sets may help to identify and address any discriminatory outcomes. Documenting all steps taken to assess and mitigate bias is also advisable. - Demand transparency from providers
Working with AI vendors who can explain how their systems work and provide evidence of fairness is preferable and, wherever possible, organisations should insist on access to information about the data and algorithms used. - Involve human oversight
The need for human intervention should never be underestimated! Ensure that AI is used to support, not replace, human decision making and that HR staff and others in similar decision-making positions understand the limitations of AI so they can spot potential issues. Sometimes automated decisions might be appealed and human review required. - Update policies and training
Recruitment and promotion policies should be revised to reflect the use of AI and the need to prevent discrimination. Providing training for staff on the risks of AI bias and the importance of fair decision making is vital to ensure a culture of inclusion and continuous improvement. - Engage with affected groups
Consult with employees, candidates, and representative groups to understand how AI systems may impact different communities and use the feedback to refine and improve AI tools and processes.
The UK government and regulators are increasingly aware of the risks posed by AI in employment. The Equality and Human Rights Commission (EHRC) has issued guidance on the use of AI in public services, and the government is considering new regulations to ensure AI is used fairly and transparently. Staying informed about developments in UK law and guidance on AI and bias is crucial to help businesses stay ahead of the curve and avoid costly litigation or reputational damage.
View more Equality Law articles from d&i Leaders
d&i Leaders is a global community of senior diversity, inclusion and HR focused professionals, looking to collaborate, network and accelerate their workplace inclusion strategy.