The European Commission has now published its legislative proposal on Artificial Intelligence. We’ll be publishing a deep dive on Freshfields Digital but highlights from the proposal include:
- new requirements for high risk AI systems, particularly where safety is affected, but also where intrusive data analytics and profiling are carried out;
- an emphasis on controls to prohibit distorting human behaviour and to improve transparency; and
- fines up to 6% of total worldwide annual turnover for non-compliance.
This legislative proposal, if it proceeds, will be a game-changer for businesses because it seeks to regulate the pervasive use of AI across a spectrum of industries and social activities. Unlike the approach in most countries so far, the EU is clearly setting out its stall to regulate all sectors and how they implement AI. Because of the broad-brush approach that the Commission has taken, virtually all systems that currently do, or which may in future, use AI would fall within scope – from personalised pricing, advertising and feed algorithms, to connected IoT systems, self-driving cars, or applications used to support recruitment and other business processes.
Together with its legislative proposal on Artificial Intelligence, the Commission also published an overarching communication explaining its approach, a new coordinated plan with Member states, a set of Q&As and a new machinery products regulation. Follow these links to see the full package:
- Proposal for a Regulation on a European approach for Artificial Intelligence and Annexes
- Communication on fostering a European approach to Artificial Intelligence
- New Coordinated Plan on AI
- Q&As
- Proposal for a Regulation on Machinery Products and Annexes
You can find more information on the implications of AI and its regulation on our AI hub.
