The European Union has unveiled the structure of its new Artificial Intelligence (AI) Agency, a pivotal part in shaping the European AI ecosystem in the foreseeable future. The Agency, constructed under the new AI Act of the EU, is expected to be operative by the end of July.
The Agency will play a dual role: managing risks associated with AI and facilitating innovation in its domain. It is designed to have global influence as many countries are devising their own AI governance strategies.
An Outline of the AI Agency
According to Tech Crunch, the Agency will be broken down into five core divisions.
The first division will be responsible for ensuring adherence to new AI laws in EU member states, maintaining a connection with national regulatory bodies, assisting them in harmonized application of legislation, and identifying and investigating potential violations. This division will have the power to impose sanctions on violators.
The second, the AI Safety division, will focus on the risks and safety posed by advanced AI models, like ChatGPT. Extra attention will be given to sophisticated General Purpose AI (GPAI) models. Their function will include identifying systemic risks of these models, developing and delivering mitigation measures, and testing them.
The third, Excellence in AI and Robotics division, will coordinate and finance scientific research and development in the field of AI, while also working to upgrade Europe’s supercomputers to support advanced model training.
AI as a Force for Good
AI for Social Good, the fourth division, will work on “AI for the benefit of society”. It will undertake development and implementation of large-scale projects with positive social impacts through the use of AI, including in areas like medical diagnostics, weather modelling, and cultural heritage restoration.
Finally, the fifth division, Innovation and policy coordination in AI, will be charged with the execution of the EU’s AI strategy. This includes monitoring trends and investments, promoting the adoption of AI technology in the industry, supporting startups, and fostering an innovative ecosystem.
Three of the five divisions will focus on fostering innovation and technology adoption, addressing concerns that new regulations could impede the growth of the sector. At the same time, two divisions will ensure compliance with law and safety requirements.
Building a robust team in the AI agency
The AI Agency is already building its team. It currently employs about 60 individuals, but aims to expand to around 140 in the coming years. The team will consist of technical experts, lawyers, economists, and other related professionals. One of the Agency’s critical tasks is to prepare guides and codes of best practice for AI developers to bridge the gap until all provisions of the Law come into full effect. It plans to publish its first guide this summer.
