The First European Commission’s Artificial Intelligence Act

Artificial Intelligence Act European Commission

The European Commission has recently announced a ground-breaking political agreement on the Artificial Intelligence Act (AI Act), marking a historic moment in the field of AI. The AI Act, proposed in April 2021, is the first-ever comprehensive legal framework on Artificial Intelligence worldwide and is set to transpose European values to a new era. The Commission’s President, Ursula von der Leyen, highlighted that AI is already transforming our everyday lives and promises enormous benefits to our economy and society.

The European Approach to Trustworthy AI

The new rules will be applied uniformly across all Member States, following a futureproof definition of AI. The Act adopts a risk-based approach, categorising AI systems into minimal risk, high risk, and unacceptable risk. Minimal risk applications such as AI-enabled recommender systems or spam filters will be free from obligations. High-risk AI systems, including certain critical infrastructures, medical devices, and certain systems used in law enforcement, will have to comply with stringent requirements. AI systems considered a clear threat to people’s fundamental rights will be banned.

The Fines for Non-Adherence

The AI Act has laid out stringent penalties for non-compliance. Companies found violating the rules will be subject to fines. The fines range from €35 million or 7% of global annual turnover (whichever is higher) for violations of banned AI applications, €15 million or 3% for violations of other obligations, and €7.5 million or 1.5% for supplying incorrect information. Small and new businesses will be able to get minor fines that are more fair if they break the AI Act. These fines show the importance of the Act and how vital it is to follow them.

The AI Act and AI for General Purpose

The AI Act establishes explicit regulations for general-purpose AI models, enhancing transparency across the entire value chain. Models that present systemic risks will carry extra legal responsibilities. These include risk management, monitoring significant events, model evaluation, and comparative testing. The Commission, businesses, academia, civil society, and other stakeholders will collaboratively develop these responsibilities into actionable conduct rules. This approach to governing general-purpose AI models demonstrates the Act’s extensive reach. Moreover, it aims to foster the responsible development of AI. 

A first for the world, the new European AI office

The European Commission is launching a new, globally unique European AI Office. This office will enforce new rules at both national and European levels. It will collaborate with national market surveillance officials to ensure correct implementation of the Act. A group of independent science experts will play a crucial role for general purpose models. They will alert about systemic risks and assist in classifying and testing the models. Together with national market surveillance authorities, this new office aims to establish a global standard for AI rules.