EU AI Act timeline for Swedish employers
5 min readPublished 2026-11-11Last reviewed 2026-11-11
Short answer
For Swedish employers, the EU AI Act's bans and AI literacy duty have applied since 2 February 2025. Rules for high-risk systems, which include AI for recruitment, promotion, task allocation and monitoring of workers, were scheduled for 2 August 2026, and the EU has proposed moving parts of them later. Start the inventory and oversight setup now.
- Prohibited practices, including emotion recognition in the workplace (from 2 Feb 2025)
- AI literacy for staff who use AI, Article 4 (from 2 Feb 2025)
What applies today and what is coming? Some rules already apply; the largest set is still being phased in.
Coming for high-risk HR uses
- Human oversight by trained staff
- Keeping logs generated by the system
- Informing workers and their representatives before use
- Following the provider's instructions for use
Empley's Readiness and Risk Control package classifies each HR AI use case and sets up oversight and logging.
Evidence
- Our assessment
Requirements summarised from public legal sources; not legal advice.
Common questions
- Is our HR chatbot high-risk?
- Usually not if it only answers policy questions. It becomes high-risk if it influences decisions about individual workers.