When does Article 14 of the EU AI Act take effect?
The EU AI Act entered into force in 2024, and its obligations apply in phases rather than all at once. The duties around high-risk AI systems, which is where Article 14's human-oversight requirement lives, arrive later in that phased schedule, as the Act comes into force. Because the exact application dates are governed by the Act's transitional provisions and remain subject to adjustment, the authoritative dates to rely on are the ones in the primary text: EUR-Lex, Regulation (EU) 2024/1689.
Why we point to the primary source for the date
Timelines in a regulation this large are staged, and proposals to adjust the timing of specific obligations have been discussed since the Act passed. Rather than restate a single date that could shift, the careful answer is to read the duty against the Act's own transitional articles and the Commission's guidance. The European Commission's regulatory framework page tracks the current state.
What to do now, regardless of the date
Treat the date as the deadline, not the starting line. Organisations that wait for the obligation to bite will be retrofitting oversight under time pressure; those that build effective human oversight now, as human judgment infrastructure, will clear the bar comfortably and run safer systems in the meantime. The law sets a floor; good practice does not wait for it. See the regulator is not your oversight strategy.
Read next
- What is Article 14 of the EU AI Act?
- Who must comply, providers or deployers?
- Back to the Article 14 hub
Sources
This page explains the law and is not legal advice. For application dates, rely on the primary text.