The deadlines at a glance
- 1 August 2024: The regulation enters into force.
- 2 February 2025: Prohibited practices and the obligation to ensure sufficient AI literacy among staff apply.
- 2 August 2025: Obligations for providers of general-purpose AI models apply.
- 2 August 2026: Transparency obligations apply, according to guidelines of the European Commission. Other parts of the regulation also become applicable from that date.
Separate deadlines apply to high-risk systems, and in November 2025 the Commission proposed changes to the timetable. Check the current status before you rely on a date.
Provider or deployer?
The AI Act distinguishes mainly between providers, who develop an AI system and place it on the market, and deployers, who use it under their own responsibility. Most mid-sized companies are deployers. Your obligations depend on which risk class a system falls into.
Where it gets tricky for industrial companies
The regulation names as high-risk, among others, AI systems in the area of employment and personnel management, for example for selecting applicants or for monitoring and evaluating the performance and behaviour of employees. Safety components of machines may also fall under stricter rules.
A simple rule for the start follows from this: do not begin with personnel evaluation, safety decisions or automatic machine control. Onagero deliberately excludes these uses in the pilot.
What an internal analysis tool typically needs
- AI literacy: The people who use the tool should know what it can and cannot do. Onagero supplies a short description of its limits with every tool.
- Transparency: Where people interact with an AI system, they should be able to recognise this. In Onagero, the model’s answers are labelled as such and point to their sources.
- Human decision: The tool prepares, a human decides. This boundary is in every Onagero pilot contract.
This is general orientation, not legal advice. The obligations depend on the specific system and its use.