An Assessment of the AI Regulation Proposed by the European Commission

TL;DR

This paper critically assesses the European Commission's proposed AI regulation, arguing it is unnecessary due to existing laws and could hinder AI development in critical sectors, potentially benefiting non-EU competitors.

Contribution

It provides a detailed analysis of the proposed regulation, highlighting its potential overreach and negative impact on AI innovation within the EU.

Findings

The regulation is deemed unnecessary due to existing legal frameworks.

It risks overregulation that could impede AI development in safety-critical sectors.

The proposal may inadvertently strengthen US and Chinese AI firms.

Abstract

In April 2021, the European Commission published a proposed regulation on AI. It intends to create a uniform legal framework for AI within the European Union (EU). In this chapter, we analyze and assess the proposal. We show that the proposed regulation is actually not needed due to existing regulations. We also argue that the proposal clearly poses the risk of overregulation. As a consequence, this would make the use or development of AI applications in safety-critical application areas, such as in healthcare, almost impossible in the EU. This would also likely further strengthen Chinese and US corporations in their technology leadership. Our assessment is based on the oral evidence we gave in May 2021 to the joint session of the European Union affairs committees of the German federal parliament and the French National Assembly.