The European Commission has marketed the European AI Office as the central hub of AI expertise. The agency will thus not have a purely regulatory character but will have broader responsibilities, such as fostering the development and responsible use of trustworthy AI while also fostering international cooperation on the rapidly evolving technology.

The office will be housed within the Directorate-General for Communication Networks, Content and Technology (DG Connect) with a team of around 100 specialists. While it will have its dedicated staff, it will also collaborate closely with existing teams within DG Connect, including those focused on the Digital Services Act (DSA) and the Digital Markets Act (DMA). Additionally, collaboration with DG Connect's supercomputing branch will provide access to powerful computing resources crucial for AI development and evaluation.

The primary task of the Office will be the support of the implementation of the AI Act. Specifically concerning general-purpose AI models, this involves:

• developing tools, methodologies and benchmarks for evaluating capabilities of general-purpose AI models, in particular for very large general-purpose AI models with systemic risks;
• monitoring the emergence of unforeseen risks stemming from general-purpose AI models, including responding to alerts from the scientific panel

More generally, this includes both general-purpose and AI systems:

• Monitoring the implementation and application of rules, in particular where the model and the system are developed by the same provider.
• Investigate infringements, including collecting complaints and alerts, assisting in the preparation of decisions of the Commission and conducting evaluations and conducting evaluations

The Office will also support the implementation of rules on prohibited AI practices. Additionally, it will assist in facilitating uniform application, and provide guidance and tools for practical implementation. The Office will contribute to standardization requests, technical support for AI sandboxes, and governance system establishment at both Union and national levels. It will also support advisory bodies, provide administrative support, and encourage Union-level codes of practice.

In terms of cooperation, the Office will engage with stakeholders, collaborate with the scientific community and AI developers, work closely with relevant Commission services, and establish cooperation with other Union bodies. Furthermore, it will play a crucial role in international cooperation on AI, advocating for responsible AI stewardship, contributing to AI regulation and governance discussions, and supporting the implementation of international agreements on AI rules.

The Office will also actively promote the development of European AI solutions. One way it will achieve this is by establishing "AI factories." These factories will provide access to the EU's powerful supercomputers, making it easier for researchers and companies to train large, complex AI models. This initiative aims to be inclusive, welcoming not just established players but also fostering innovation by making these resources available to Europe's burgeoning AI startups,