The European Union is putting a formal operating structure behind its AI Act. The bloc has outlined a five-unit AI Office that will help enforce the new risk-based regulatory framework while also trying to build confidence, investment and adoption around artificial intelligence.
The AI Office will take effect on June 16. The AI Act is expected to enter into force before the end of July, after final approval by EU lawmakers last week.
A new center for AI governance
The AI Office is designed to sit at the center of the EU's AI governance plan. Its work is not limited to policing risks. It is also meant to help shape the European AI ecosystem in the coming years, including through research support, international engagement and policy coordination.
That dual mission matters. The EU has moved quickly to create a rulebook for artificial intelligence, and some homegrown AI developers have complained that the approach could slow innovation. By assigning three of the five units broadly to uptake, investment and ecosystem building, the Commission appears to be signaling that regulation and growth are meant to move together.
The bloc also wants the AI Office to matter beyond Europe. Many countries and jurisdictions are still deciding how to approach AI governance, and the EU hopes its model can carry influence internationally.
Compliance and AI safety form the enforcement core
Two units will carry the most direct regulatory responsibilities.
The first will focus on regulation and compliance. Its work includes liaising with EU Member States to support harmonized application and enforcement of the AI Act. According to the Commission, the unit will contribute to investigations, possible infringements and administering sanctions.
That role is supportive as well as operational. The AI Act will also establish EU country-level governance bodies, and the AI Office is expected to help connect those national enforcement efforts into a more consistent system.
The second enforcement-focused unit will handle AI Safety. The Commission said this team will focus on the identification of systemic risks from very capable general-purpose models, possible mitigation measures, and evaluation and testing approaches.
General-purpose models, or GPAIs, refer to the recent wave of generative AI technologies, including foundational models behind tools such as ChatGPT. The source notes that the EU is especially focused on GPAIs with so-called systemic risk, which the law defines as models trained above a certain compute threshold.
The AI Office will directly enforce the AI Act's rules for GPAIs. That means relevant units are expected to carry out testing and evaluation, and to use powers to request information from AI giants when needed for oversight.
The compliance work will also include producing templates that GPAIs are expected to use. One example is a template for summarizing any copyrighted material used to train their models.
Research, public good and innovation get dedicated units
The remaining three units are focused on developing the AI ecosystem rather than only controlling risk.
One unit will cover Excellence in AI and Robotics. Its remit includes supporting and funding AI R&D. The Commission said this team will coordinate with the previously announced GenAI4EU initiative, which aims to stimulate the development and uptake of generative AI models, including by upgrading Europe's network of supercomputers to support model training.
Another unit will focus on AI for Social Good. The Commission said it will design and implement the Office's international engagement for major projects where AI could have a positive societal impact. The examples given include weather modelling, cancer diagnoses and digital twins for artistic reconstruction.
This direction had already begun to take shape. Back in April, the EU announced a planned AI collaboration with the U.S. on AI safety and risk research, which would also include joint work on AI uses for the public good.
The fifth unit will handle AI Innovation and Policy Coordination. Its job is to ensure execution of the bloc's AI strategy. The Commission described this as including monitoring trends and investment, stimulating AI uptake through a network of European Digital Innovation Hubs and the establishment of AI Factories, and supporting regulatory sandboxes and real-world testing.
Leadership and staffing are still taking shape
The Commission has already appointed several leaders for the AI Office, although not every key position is filled.
Confirmed appointments are:
- Lucilla Sioli, head of AI Office
- Kilian Gross, head of the Regulation & Compliance unit
- Cecile Huet, Excellence in AI and Robotics Unit
- Martin Bailey, AI for Societal Good Unit
- Malgorzata Nikowska, AI Innovation and Policy Coordination Unit
The head of the AI Safety unit has not yet been named. A lead scientific advisor role is also vacant.
The AI Office was established by a Commission decision back in January and began preparatory work in late February, including deciding its structure. It sits within the EU's digital department, DG Connect, which is currently headed by internal market commissioner, Thierry Breton.
Its eventual headcount will be more than 140 people, including technical staff, lawyers, political scientists and economists. On Wednesday the EU said some 60 staff have been put in place so far. Hiring is expected to increase over the next couple of years as the law is implemented and becomes fully operational.
What comes next as the AI Act phases in
The AI Act will not apply all at once. It takes a phased approach, with some provisions set to apply six months after the law comes into force and others getting a longer lead-in of a year or more.
That makes the AI Office's near-term work important. One key task will be drawing up Codes of Practice and best practices for AI developers. The EU wants these to serve as a stop-gap while the legal rulebook is phased in.
A Commission official said the Code is expected to launch soon, once the AI Act enters into force later this summer.
The AI Office will also work with several other bodies created by the AI Act. These include the European Artificial Intelligence Board, made up of representatives from Member States; a scientific panel of independent experts; and a broader advisory forum with stakeholders including industry, startups and SMEs, academia, think tanks and civil society.
The Commission noted that the first meeting of the AI Board should take place by the end of June. It also said the AI Office is preparing guidelines on the AI system definition and on prohibitions, both due six months after the entry into force of the AI Act. Codes of practice for obligations on general-purpose AI models are due 9 months after entry into force.
The structure now outlined shows how the EU plans to turn the AI Act from a legal framework into a working governance system. Its success will depend not only on enforcement powers, but also on whether the Office can coordinate national authorities, engage developers and keep pace with general-purpose AI models as the rules phase in.