The European Commission has published its proposal for a new, dedicated Regulation on Artificial Intelligence. At the core of the proposal lies a risk-based regulatory approach, whereby certain AI practices are to be prohibited, others are to be labelled “high-risk” with correspondingly strict requirements, while other AI systems which pose “limited risk” and “minimal risk” will have to adhere to basic transparency requirements or can operate without facing additional rules, respectively.
The Commission’s aim is to balance the need to protect European citizens, an approach inspired by the General Data Protection Regulation (GDPR), while enabling (or at least not undermining) the uptake of AI systems needed to future-proof European industry.
While many further discussions are needed to establish whether the proposal achieves the right balance and constitutes an effective and feasible way forward, the text certainly provides a basis for discussions in Europe, as well as globally.
Hume Brophy has looked at the proposal in depth and outlines what you need to know.