The European Commission published a comprehensive proposal to regulate the use of Artificial Intelligence (AI) technologies on April 21st 2021.
I have recorded a new podcast episode with Nicklas Berild Lundblad where we share some initial reactions to and thoughts about this proposal.
A particular concern of mine is that there should be an open discussion about the political intent in respect of specific uses of AI to help us understand what politicians in the EU want and expect to happen when the regulation is adopted.
Some areas are covered in the proposal in quite a lot of detail such as the deployment of AI that uses biometric data in public places for law enforcement purposes.
Other uses that are of considerable public interest, such as AI-driven ad targeting and content moderation are not legislated for specifically but would rather be subject to general provisions.
These are of course early days in the evolution of the legislation and there will be several years of debate before it reaches its final form, but I hope we can discuss concrete examples sooner rather than later.
You certainly cannot cover every eventuality in a broad regulatory instrument and this is not a call for everything to be set out in the legal text in painful detail, but there is an opportunity to clarify and test how it is hoped the law will be applied.
Doing this now at least for obvious and important uses of the technology is better, in my opinion, than leaving too much open for different and conflicting interpretations that will have to be settled by courts down the line.
I am sure I will come back this will more written analysis later but for now our recorded conversation may be useful for anyone who is also looking at this proposal and forming their own views of it.