EU Rules on AI – some thoughts
The European Comission has presented their proposal on AI regulation last month. It is the first of its kind. I think there is no question that Europe has to find its own way to deal with AI. Trustworthy and human centric AI is a very good approach without question.
But the devil is in the details and I agree with Dragoș Tudorache, Romanian MEP and chair of the European Parliament’s committee on AI, that this proposal is complex and has to be reviewed in detail.
A few thoughts
1/ We have to focus on innovative solutions fist – and risk second. Otherwise Europe will not become a leading player in the AI world. We have to give innovation a chance
2/ When regulation is needed it has to be clear (learning from GDPR) – even if you do not have a lawyer (or legal department)
3/ The more risk classes we introduce the more bureaucracy we will get. We should limit this to the highest risk areas.
4/ And before creating new regulation (and regulatory bodies) we always should try to adopt current laws as a default. In many areas we already have laws and regulations in place which also apply to AI solutions or can be adapted.