The European Commission’s plan to hire some 100 staff members for its new AI Office and finance it with a reshuffled budget from the Digital Europe Program has raised questions from EU member states, according to minutes of a meeting seen by Euronews.
The AI Office, announced by the commission last week (24 January) and set to enter into force on 21 February, will be tasked with supervising the rules for general-purpose AI systems under the upcoming AI Act. It will also function as a central coordination body for AI policy at EU level, coordinating with other commission departments, EU agencies, companies and the 27 EU Member States.
However, the executive’s plan to finance its establishment through a reshuffle of the budget of the Digital Europe Program, which is funding designed to go to projects related to cybersecurity and supercomputing, sparked questions from the 27 capitals during a meeting between the commission and diplomats held earlier this week.
Several countries including Denmark, Finland and Sweden asked for more information on how the finances will be redistributed. In addition, they want to be reassured that the new office will be adequately staffed.
The commission said in a response that additional financing was not counted on within the Union’ multi-annual budget plan (MFF), which covers the period 2021-2027.
The AI Office will fall under the commissions’ digital unit and will therefore be subject to the management plan of DG Connect. According to the minutes, some 100 staff members will be needed. The plan is to attract 80 employees from elsewhere and to transfer 20 people from within the commission to the new office.
As with the AI Office, the executive is also seeking employees to work on incoming platform regulation. In the past months, it sent out several calls for applications to build up a team within DG Connect, combining oversight on the Digital Services act and Digital Markets Act rules, with over 100 full time staff.
France, one of the countries that has been sceptical about the technical details of the AI Act, expressed concerns about the commission’s timeframe, as there is no official deal on the AI rulebook yet. Negotiators of the EU institutions are about to close a deal on the technical details on Europe’s plan to regulate high-risk AI applications, after a political agreement in December.
If the final text is greenlighted by ambassadors tomorrow (2 February), the European Parliament is expected to vote in its Internal Market and Civil Liberties committees mid-February, and in plenary in March or April. After that, the act is expected to enter into force later this year and includes an implementation period of up to 36 months.
In addition to the AI Office, and national regulators, three other supervision and enforcement bodies will be set up.
The European Artificial Intelligence Board will include representatives of each member state to assist the commission and EU countries and share regulatory experience. A separate Advisory Forum will consist of industry representatives, SMEs and academia to give technical input on the AI Act. And finally, a Scientific Panel, a pool of independent scientific and technical experts, will support the AI Office in the implementation and enforcement when it comes to General Purpose AI models.