Recruitment and workforce-management tools need a use-case review. The revised timetable should not be confused with the rules already applicable.
What the official timeline says
The Commission’s AI Act Service Desk identifies 2 August 2026 as the start of the majority of the Act’s rules, including transparency requirements. Following the Digital Omnibus amendments, Annex III high-risk-system rules are scheduled for 2 December 2027, while high-risk systems embedded in regulated products follow on 2 August 2028. Earlier AI literacy and prohibited-practice provisions remain relevant.
What this means for employers
Do not describe all employment-related high-risk duties as having started in August 2026. Equally, a later high-risk milestone does not make a proposed HR system lawful today. Existing data-protection, discrimination and employee-representation rules still need to be assessed.
A practical inventory
List tools used for applications, interviews, scheduling, performance and promotion. For each one, identify the provider, the data used, the output, the human decision-maker and the consequences for individuals. Ask vendors for the documents that describe the actual configured product, not only marketing assurances.
Before deployment
Agree access, retention, correction and escalation procedures. Review works council involvement and whether the human reviewer can meaningfully challenge the output. Keep the implementation schedule under review against the consolidated legislation and official guidance.
Sources and further reading
Published: 8 November 2026. Related legal advice · Contact the firm