Luca Bertuzzi

The @eu_eeas is working on an 'AI in Security and Defense Implementation Roadmap' to coordinate AI initiatives in foreign policy and defense. The focus is on AI-enabled threats, capabilities, global governance, & cooperation with partners. https://t.co/vxcQ9H2CTL

🚨Together with the AI Continent Action Plan, the Commission will open a public consultation next week on how to simplify AI Act compliance. The plan includes ideas like data labs, a Resource for AI Science in Europe & pathways for foreign AI talents. https://t.co/s9sk8QCIu9

EU countries are taking different paths on AI governance approaches. According to a questionnaire's results, common issues are emerging in the AI Act implementation, but there is also a growing divergence on critical aspects, prompting calls for guidance. https://t.co/C6Eqx4uHJ6

As the AI Act's deadlines are fast approaching, its applicability is raising important legal questions, notably concerning the development phase, potential extraterritorial effects, key concepts like market placements, and the scope of the R&D exemption. https://t.co/kcPOZmhbCW

The Commission is buying time on whether it will present a legislative initiative on algorithmic management. Yesterday, it published an already outdated report, one and a half years after it was concluded, and in January, it outsourced another study. https://t.co/eslOesdAbA

EXCLUSIVE: Very large online platforms paid €58 mil in supervisory fees under the #DSA in 2025, according to a draft report seen by @mlexclusive. The report breaks down last year’s expenditures and states that the costs were underestimated by €0.5 mil. https://t.co/I6IXjXoJoZ

Poland opted for an integrated supervisory approach to AI governance, setting up a new AI Commission composed of existing regulators. However, this new body raised questions about the involvement of independent authorities, most notably the DPA. https://t.co/KMs5cXuE4M

The D9+ countries are set to call for reviewing EU digital legislation, deregulating 'where possible' and finding ways to foster private investments in European digital technologies, according to a draft ministerial declaration seen by @mlexclusive. https://t.co/R8ldBz7SVW

Germany's next government might ask for a revision of the AI Act, according to leaked drafts of the coalition papers. While this revision clause is still tbc, Germany seems set to implement the AI law in an innovation-friendly manner. https://t.co/6ktJix7efA

The European Commission updated EU member states yesterday on the upcoming initiatives to implement the AI Act, notably on the guidelines on the rules for general-purpose AI, the requirements for high-risk AI systems, and the Advisory Forum. https://t.co/yf1ZnZDGOy

NEW: The EU legislators who directly negotiated the AI Act sent a letter to the Commission today raising 'great concern' on the GPAI code of practice, saying that the risk hierarchy introduced in the third draft is against the law. https://t.co/2He5vH3lup

Last week, the Italian Senate adopted a draft AI law with an amendment obliging all AI systems meant to be used in the public domain to be based on servers located in Italian territory. If confirmed, it would have massive legal and economic implications. https://t.co/JgWJmJxoVz

Those hoping for the Commission to give more details on the AI Act's template for model training data were disappointed, as only aggregate data on the feedback was presented. The discussions will continue, incl. at the political level. The timeline is Q2. https://t.co/DRfDG0KW39