In April 2021, the European Commission presented its long-awaited proposal to regulate the use of artificial intelligence (AI) in Europe. The proposal was accompanied by a new coordinated plan for the Commission and the 27 EU member states to increase the uptake of AI.

The proposed AIA follows a risk-based approach: the higher the risk, the tighter the restrictions. At the heart of this approach are high-risk AI systems, i.e. AI systems which create an adverse impact on people's safety or their fundamental rights. The proposed regulation would subject high-risk AI systems to mandatory requirements, an ex-ante conformity assessment and post-market monitoring.

In addition, the new AI rulebook would:

• ban certain AI systems which present an “unacceptable risk” (such as AI systems which use subliminal techniques to distort a person’s behaviour);

• lay down harmonised transparency rules for AI systems intended to interact with natural persons (such as emotion recognition systems);
• set rules on post market monitoring and measures for non-compliance, where failure to comply with the AIA could result in significant administrative fines of up to €30m or 6% of a company's annual worldwide turnover; and
• put in place governance systems at a national and an EU level (through a new European Artificial Intelligence Board).