The European Union’s AI rules have entered a new phase. The AI Omnibus came into force on 27 July 2026 and changed parts of the AI Act. The Act’s general application date is 2 August 2026, while some rules follow later dates.
For certain high-risk AI systems used in areas such as hiring, education, and access to essential services, the main rules will apply from 2 December 2027. These requirements cover topics such as risk checks, records, clear information, human oversight, and security.
High-risk AI built into regulated products, including machinery, toys, and lifts, has a later date: 2 August 2028. This gives manufacturers more time to prepare, but it does not remove the rules.
The update also extends some simpler requirements for small businesses to small mid-sized companies. It expands supervised testing spaces, called regulatory sandboxes, and gives the EU AI Office a wider role in overseeing some AI systems.
The main point is that the law has a timetable, not one single start date for every duty. Some AI Act requirements already apply, while the high-risk rules described above begin later. Companies should identify what their AI system does and check which legal category and deadline apply.
Sources:
- European Commission Official regulation
- European Commission
- AI Act overview Regulation (EU) 2026/1744