FCA authorised · FRN 724952 0117 325 0027Quote & buy →
Aipedia: HR consultants
AI-scored video interviews and online assessment games profile candidates, so automated decision rules and the Equality Act both apply. HR advisers who recommend or configure these tools need to check bias, transparency and adjustments before a client goes live.
Part of: AI in HR consultants
In short
AI video interview and game-based assessment tools infer traits from how candidates speak, look or play. The ICO's 2024 audit found recruitment tools inferring gender and ethnicity, and its draft ADM guidance warns decision-support tools often make significant decisions. Rejection without meaningful human review triggers Article 22C safeguards. Scoring can disadvantage disabled, neurodivergent or ethnic minority candidates, and Manjang v Uber Eats shows how facial checks can lead to claims. Advisers should check bias testing, adjustments, human review, DPIAs and any vendor conflicts.
Some recruitment platforms ask candidates to record answers to set questions, then use AI to score the recordings on content, speech patterns or behaviour. Others use online games or psychometric tasks and score candidates against a profile of successful employees. HR consultants are often asked which tool a client should buy, how to configure it and how to weigh its scores against other evidence.
These tools profile candidates. They infer traits from data, and that inference is where data protection and discrimination risk sit.
The ICO's 2024 audit of AI recruitment tool providers found tools that let recruiters filter by protected characteristic, inferred gender or ethnicity from names, and kept excessive data indefinitely. Almost 300 recommendations followed, and the ICO said recruiters should use its key questions when buying tools. Its draft ADM guidance says "decision-support" tools often turn out to make significant decisions, and that reviewers must understand a system's limits.
Under UK GDPR Article 22A, the extent of profiling must be considered when deciding whether a decision is solely automated, and a rejection is a significant decision. If no human meaningfully reviews the result, Article 22C safeguards apply. The wider lawful bases for automated decisions do not extend to special category data, which matters if a tool infers health, disability or ethnicity.
Video and behavioural scoring can disadvantage people whose speech, expression, eye contact or response time differs for reasons connected to disability, neurodiversity, race or age. Facial analysis is a particular concern: in Manjang v Uber Eats, facial recognition checks repeatedly failed to verify a Black courier, an automated process removed him, and the EHRC-backed claim settled in 2024. The EHRC's message was to be transparent about AI, explain decisions, offer a way to challenge them and guard against discrimination. Employers also have a duty to make reasonable adjustments for disabled candidates, which a fixed automated process may not allow.
A further problem is explainability. If the vendor cannot say why a candidate scored poorly, the employer cannot explain the decision to the candidate, cannot answer a tribunal's questions and cannot show that a human reviewer understood the system's limits.
Illustrative: a consultant recommends an AI video interview platform to a client and helps set the pass mark. An autistic candidate scores poorly on "engagement" measures, is rejected without human review and was not offered an alternative format. The candidate brings disability discrimination and data protection complaints. The client argues its consultant did not warn it about adjustments or automated rejection.
If the client pursues you for negligent advice on choosing or configuring the tool, PI usually responds, subject to the wording, including any limits on discrimination-related losses or data protection claims. If your firm resells or licenses the tool, the risk looks more like a technology product, and you should discuss that with your broker. See AI bias and discrimination claims.
Apex arranges PI for HR consultants who advise on recruitment technology, and a broker can check how your wording handles vendor selection advice. See also AI CV screening, the ICO rules explained and AI in HR consulting.
They are not banned, but they involve profiling, so data protection rules on automated decisions apply, and they must not discriminate. A DPIA, bias testing, meaningful human review and a route to adjustments are expected.
If a rejection is a solely automated significant decision, Article 22C requires the candidate to be informed and able to challenge it. The ICO also wants employers to be more transparent about automated recruitment generally.
Automated scoring can disadvantage candidates whose communication differs because of disability. Employers must consider reasonable adjustments, such as an alternative format, and HR advisers should check the tool allows them before recommending it.
Yes. The CIPD Code requires members to identify and disclose conflicts. A referral arrangement could also be used to challenge the independence of your recommendation if the tool later causes a claim.
Tell us about your firm and the work you do. A broker will review how your current PI treats AI-assisted work, or arrange new cover. No fee for a quote. Or call 0117 325 0027.
Start a PI quote Request a callbackApex Insurance Brokers Limited is authorised and regulated by the Financial Conduct Authority. Registered in England and Wales, company number 07014570. This page is general information about how AI is used in the profession and how insurance may respond. It is not legal, regulatory or insurance advice on your circumstances, and it does not guarantee that cover will be available or on what terms.
