Do your recruitment practices avoid AI-related risks?

Why should I read this?

The Department for Science, Innovation and Technology has published guidance on ‘Responsible AI in Recruitment’ (DSIT Guidance), to seek to address and mitigate some of the risks from the use of artificial intelligence tools in this area.

While the DSIT Guidance is not mandatory, its contents implement the Government’s AI regulatory principles which are enforced by a number of regulators, including the equality and human rights and data protection watchdogs. All employers should therefore take account of the DSIT Guidance in their recruitment practices.

In this briefing, we provide an overview of the DSIT Guidance and suggest some of the practical actions that employers should consider taking.

What do I need to know?

The UK government's White Paper, "A pro-innovation approach to AI", published in March 2023, outlined the UK's AI regulatory framework. The stated underlying aim is to create a cross-sectoral, principles-based framework to build trust in AI and in which innovation can thrive. The five principles that underpin that framework, and which are interwoven throughout the new DSIT Guidance, are safety, security, and robustness; appropriate transparency and ‘explainability’; fairness; accountability and governance; and contestability and redress.

Noting the novel risks posed by the use of AI-enabled tools at each stage of the hiring process, the DSIT Guidance highlights the core considerations for organisations when procuring and deploying AI in recruitment to take account of the framework principles, including suggested questions and AI assurance mechanisms. The guidance also suggests some questions that could be asked during the procurement of new AI tools to verify suppliers’ claims about those tools.

The DSIT Guidance outlines what organisations should consider when procuring an AI system, including the purpose and desired output of that system, how it will integrate into existing processes, and whether employees will need training or additional resources to use the system. In addition, the DSIT Guidance highlights that the use of AI systems can bring novel risks of discrimination, and therefore the importance of ensuring that new barriers to applicants with protected characteristics are considered so that they can be avoided.

Data protection is also a focus of the DSIT Guidance, highlighting that for some automated processes, there may be a right for data subjects not to be subject to a decision based solely on that processing. In that event, certain safeguards to protect the data subject’s rights and freedoms may be required, including some human involvement in that process. The Guidance further highlights that the completion of a data protection impact assessment is required for all development and deployment of AI systems that involve personal data, as well as making clear that the use of AI systems in recruitment should be clearly signposted to applicants.

What should I do next?

Despite the non-mandatory nature of the guidance, taking account of the recommended actions will give organisations an important head start as regulatory scrutiny increases in the area of AI. Particularly in light of the regular use of AI technology in recruitment, this is an obvious area where understanding, assessing and managing the risks can make an important difference.

The DSIT Guidance suggests that employers should create an AI governance framework, including assigning who is accountable for the AI system, any training requirements, the communication strategy around it, and regular evaluation and monitoring to assess the risks. That evaluation should be undertaken at the procurement stage of any AI tool, as well as at regular intervals during its deployment. The exact nature of the assurance process will depend on the particular tool and its intended use, however this may include bias audits, performance testing, risk assessments, user feedback systems or the use of model cards.

As a minimum, employers should therefore consider:

• putting in place an AI governance framework
• implementing processes to ensure appropriate information-gathering pre-procurement
• performance testing and monitoring
• identifying and addressing training needs
• creating an effective communication strategy to ensure transparency in deployment
• undertaking ongoing risk and impact assessments

Aside from the topic of the use of AI in recruitment, organisations can expect to see guidance published in other areas of AI use too, and should continue to monitor developments in this respect.

Looking ahead

The Government is expecting individual regulators to take swift action to implement the AI regulatory framework, giving them until 30 April 2024 to publish their strategic approaches. The publication of these plans will provide further insight into regulators' strategic direction and forthcoming guidance and initiatives in this area.

Further reading

For further information on this topic, please see the Government publications: DSIT Responsible AI in Recruitment guide, the White Paper - AI regulation: a pro-innovation approach and the consultation outcome response.

Please also see our previous briefings on the topic of the use of AI in employment: AI at work: addressing the employment risks to realise the rewards and Global AI at work - Regulating responsible AI use.

How we can help

Our teams have significant experience of supporting employers to steer through the legal, regulatory and practical implications associated with the implementation and use and AI technologies in the workplace, including around employee relations and engagement.

Please contact any of our team should you require advice or assistance.