AI is changing the way we work, and recruitment of staff is one area where all kinds of businesses are hoping to reap the benefits of the developing technology.
But the exciting opportunities which AI presents, also bring risks – legal and reputational – which businesses need to manage.
We spoke to Robert Newry at Arctic Shores (AI thought leader and LegalEdge client) to ask for his top tips on using AI in the recruitment process and what pitfalls to avoid. Here’s what he told us:
1. Make sure your model is fit for purpose:
Efficiency is not the same thing as effectiveness (quality of hire). Businesses talk about “improving efficiencies” with AI, imagining that automation is the magic bullet, but it is no good being efficient if you don’t end up with better hires – you’ll just be right back to square one, without the person your business needs.
2. Identify what “good” looks like for your organisation/the role
Your process should be designed with this in mind from the outset (too often the job description is outdated or the selection process is generic). Identify the skills and capabilities you need both for today and in the future at the outset – this will help ensure consistency in the process as well as relevant screening stages.
3. Don’t overlook the data protection/security implications of what you are doing
Asking ChatGPT to rank CVs is a no-no! Recruitment inevitably involves personal and sensitive data so you should never use an open-source, public model for such a task. Furthermore, it’s not consistent or reliable however credible it may seem at first sight.
4. Ask yourself how you will ensure fairness
For example, you should be transparent and upfront with candidates at the start of the process – tell them if you are using AI, where you are using it and how you are using it. This will help to signal that you respect the effort of candidates and will be an investment in trust for the future.
5. Human oversight is crucial, consider carefully where this is best inserted into your process
You still need a mixture of machine and human input despite the often overwhelming volume of applications. For example, if you are operating in regulated sector such as financial services, insurance or medicine there are compliance requirements for which AI may not be suitable in selection, and where human intervention is required.
6. Don’t ignore the potential for AI bias, how can your business be sure it is not discriminating against candidates?
Bias is a real risk due to the nature of AI and the way it functions. Independent validation of your model can help to evidence that this is an acknowledged risk which has been reviewed and, as far as possible, mitigated. Candidates may bring a claim for unlawful discrimination in the employment tribunal on all the usual grounds (sex, race, discrimination, age etc.) and validation could be a helpful defence to any such claim.
7. Look out for the AI enabled candidate
AI is just as freely available to candidates as it is to recruiters. Businesses are routinely having to deal with many, many more applications as platforms introduce “one click” applications and individuals are able to make many more applications, with the help of AI. Reliable screening tools can make a real difference and an in person interview therefore takes on an even greater importance as a tool identifying the right candidate.
And finally,
With the Employment Rights Bill moving closer, the costs of a poor recruitment decision are going to go up. We expect staff will be protected from unfair dismissal from day one (see our recent blog: Employment law changes), so the pressure to get the right person first time will increase. While AI presents exciting opportunities, businesses need to carefully manage the integration of AI into their recruitment processes to avoid costly mistakes.
Can we help you?
We can help you review and amend your recruitment policies and processes and help implement practical changes to ensure you remain compliant and mitigate legal risk. We also offer training to help you get to grips with new legal requirements.
In the age of Generative AI and shifting skill demands, Arctic Shores empowers organisations to look beyond CVs and uncover true human potential. Its task-based assessments generate far richer insights than traditional methods, using validated models that help widen talent pools, counter bias, and deliver diverse, high-potential candidates in any economic climate. By combining this next-generation technology with careful oversight and legal compliance, businesses with volume recruitment roles
can build a fairer, more resilient workforce while avoiding the risks that come with poor recruitment decisions. Let us know if you’d like to chat to them.
