AI may help recruitment teams organise information and prepare communication. Candidate data and decisions need particular care because mistakes or bias can affect people’s opportunities.
Possible uses
An approved tool might extract information from a CV, help organise notes or draft a follow-up message. Check that the extracted details are correct and relevant before using them.
Screening and shortlisting
A model can overlook relevant experience or produce biased rankings. Do not assume AI removes bias or that a human merely approving its suggestion makes a process fair. Test for unfair outcomes, make review meaningful and assess the rules that apply to decisions affecting candidates.
Privacy and transparency
Tell candidates how their information is used, check what the supplier does with it and limit access and retention. Record or transcribe interviews only where appropriate and after providing the required information. Obtain specialist advice on data protection and employment obligations.
Start carefully
Consider starting with a low-risk administrative task, then measure accuracy and the effect on candidates before expanding. Give candidates a way to raise concerns or correct their information.
Start with a reviewable task
A team might test drafting an acknowledgement from an approved, non-sensitive example. A recruiter checks the recipient and wording before sending. If a tool extracts candidate information, check it against the source and provide a way to correct mistakes.
Before using AI to rank or make decisions about candidates, involve appropriate employment and privacy specialists. Assess the effect on applicants, the information provided to them and whether review is meaningful rather than automatic approval.
The ICO’s recruitment AI guidance sets out questions to ask a provider.
