UK publishes responsible AI recruitment guidance

GOV.UK has published new guidance for employers and recruiters on the responsible use of artificial intelligence in hiring, setting out practical steps for organisations that deploy automated tools for CV filtering, candidate scoring and other selection decisions.

The guidance, published on GOV.UK, was developed with input from the Information Commissioner's Office and the Equality and Human Rights Commission and draws on best-practice material from ACAS and the CIPD. It addresses core topics employers face when introducing automated recruitment tools, including transparency for candidates, bias mitigation, human oversight and legal duties under the Equality Act 2010 and UK GDPR.

At its centre the document frames automated hiring tools as lawful only when organisations can demonstrate appropriate risk assessment and controls. Employers are advised to carry out audits of systems before and during deployment, document decision-making and keep records of testing and monitoring. The guidance recommends data protection impact assessments where processing is likely to pose a high risk to individuals, and suggests steps for reducing discrimination risks — for example by reviewing training data, tracking disparate impact on protected groups and introducing human checks on automated recommendations.

On transparency, the guidance asks employers to provide clear information to candidates about where automation is used and what it does: how an algorithm influences shortlisting, what data is used and how applicants can request human review. It also encourages employers to set roles for human oversight that go beyond token sign-off, so that staff can interrogate and override automated outputs when necessary.

The publication places particular emphasis on aligning automated hiring with duties under the Equality Act 2010. Employers are reminded they must avoid indirect discrimination and consider reasonable adjustments for disabled candidates where algorithmic processes could disadvantage them. The guidance also explains how UK GDPR obligations — such as lawful bases for processing and transparency requirements — intersect with recruitment systems that rely on personal data.

The launch reflects a wider regulatory push in the UK to bring more practical advice to employers as talent-technology adoption accelerates. Regulators including the ICO and the EHRC have increasingly signalled they will expect organisations to show proactive governance of automated decision‑making; this guidance offers a relatively detailed playbook that HR teams and procurement leads can use in vendor selection, testing and on-going oversight.

What the guidance does not do is create new legal thresholds or introduce mandatory certification for recruitment algorithms. It stops short of prescribing specific accuracy rates, audit frequencies or independent certification schemes, and it does not list approved vendors or tools. The document also offers limited practical support for small and medium-sized employers on resource implications, and it contains no enforcement timetable describing how or when regulators will test compliance in the field.

For HR leaders, the guidance tightens expectations: vendors and internal teams will increasingly be judged on documented audits, demonstrable human oversight and candidate-facing transparency. As algorithmic hiring becomes routine, employers will need to bake the new checks into procurement, training and record-keeping to show they meet equality and data-protection duties. The guidance is likely to influence contract terms with suppliers and could be the baseline regulators point to when assessing future complaints or investigations.