SHL

Company overview

SHL (legal entity SHL Group Ltd.), founded in 1977 and headquartered in the UK with major US operations, is one of the world's largest psychometric and talent-assessment firms, serving most of the FTSE 100 and over half the Fortune Global 500. Its portfolio spans cognitive ability, personality, behavioral, and situational-judgement tests plus AI-scored products acquired largely through its 2019 acquisition of Aspiring Minds, including Smart Interview (video interviewing), Automata (coding), SVAR (spoken language), WriteX (writing), and conversational-chat simulations. SHL grounds its products in industrial/organizational psychology and publishes extensive validity and fairness research, positioning itself as an enterprise-grade, science-led assessment vendor.

Regulatory exposure

SHL's AI-scored assessments score, rank, and recommend candidates, placing them squarely in scope as high-risk AI under EU AI Act Annex III and as AEDTs under NYC Local Law 144. SHL has substantial public-facing material for NYC LL 144 (a detailed FAQ and a US regulatory-compliance page) and grounds validation in the EEOC's AI technical-assistance guidance and the Uniform Guidelines, but its NYC bias audits are conducted by an independent auditor on a per-customer, fee-based, NDA-gated basis (hosted in an internal repository) rather than published openly. SHL maintains a public AI Policy and best-practice whitepaper addressing explainability, fairness, and human oversight, and holds ISO 27001/27018/27701 plus CSA STAR for security/privacy, but publishes no ISO 42001, no EU AI Act-specific deployer guidance, no Article 27 FRIA template, and no independent third-party algorithmic-fairness audit (e.g., Warden AI, Holistic AI, BABL AI).

Path to a higher score

SHL's strongest lever is to convert its already-detailed internal governance into public, downloadable artifacts: publish at least one independent NYC LL 144 bias-audit report (with impact ratios and auditor named) and commit to consecutive yearly public audits; commission an independent third-party AI assurance/audit and publish the result. It should add explicit EU AI Act content, an Article 27 FRIA template or deployer-obligation guidance for high-risk recruitment use, pursue ISO 42001, and publish per-product model cards/instructions-for-use plus a public security disclosure channel, status page, and model-update changelog to evidence post-market monitoring.