EU locks AI hiring rules
- The EU confirmed an August 2026 deadline for AI hiring rules that regulate high‑risk HR systems. - Fines for non‑compliance can reach €15 million under the implementation update cited today. - That timeline makes recruitment and performance tooling potential regulated high‑risk systems that need audits and documentation (asanify.com).
The European Union has fixed August 2, 2026 as the date when its main Artificial Intelligence Act rules hit hiring and worker-management systems classed as high risk. (digital-strategy.ec.europa.eu) The European Commission’s AI Act timeline says that on August 2, 2026, “the majority of rules” start to apply and enforcement begins, including the Annex III rules for high-risk systems. Annex III covers AI used in employment, worker management, and access to self-employment. (ai-act-service-desk.ec.europa.eu) (eur-lex.europa.eu) That category includes tools used to recruit or screen applicants, make decisions on promotion or termination, allocate tasks, and monitor or evaluate workers. The law treats those uses as high risk because they can affect pay, jobs, and access to work. (eur-lex.europa.eu) (digital-strategy.ec.europa.eu) For employers and software vendors, “high risk” means paperwork and controls, not an outright ban. Providers and deployers must be ready for requirements on data governance, human oversight, record-keeping, and instructions for use when those rules take effect. (digital-strategy.ec.europa.eu) (ai-act-service-desk.ec.europa.eu) The penalty ceiling is also set in the law. Article 99 allows fines of up to €15 million, or 3% of worldwide annual turnover, for a range of violations outside the Act’s outright bans, including deployer obligations and transparency duties. (ai-act-service-desk.ec.europa.eu) The AI Act is not landing all at once. It entered into force on August 1, 2024; bans on prohibited practices started applying on February 2, 2025; general-purpose AI rules started on August 2, 2025; and the broader high-risk regime arrives on August 2, 2026. (commission.europa.eu) (ai-act-service-desk.ec.europa.eu) Some workplace AI uses are banned before that broader deadline. The Commission says emotion-recognition systems in workplaces are among the prohibited practices already covered by the Act. (digital-strategy.ec.europa.eu 1) (digital-strategy.ec.europa.eu 2) Deployers of high-risk systems must also use them according to the provider’s instructions, assign human oversight, keep automatically generated logs for at least six months, and inform workers or their representatives when the system is used in the workplace. (ai-act-service-desk.ec.europa.eu) A separate fundamental-rights impact assessment is required only for certain deployers, including public bodies and some private operators, before first use. That assessment must describe the system’s purpose, affected groups, risks, oversight, and mitigation steps. (ai-act-service-desk.ec.europa.eu) The practical effect is that companies selling résumé filters, interview scoring, productivity scoring, or employee monitoring into Europe now have a fixed compliance date. By August 2, 2026, the question is no longer whether the rules are coming, but whether the system can pass them. (ai-act-service-desk.ec.europa.eu) (eur-lex.europa.eu)
