I’m just back from Copenhagen where we ran a roundtable on AI assurance, asking how we can best use tools to reduce risk and address compliance with upcoming regulations, particularly the European AI Act. The speaker details are in the link, and there will be a full video too. We had about 25 people in the room, and another 15 or so online throughout the two hour meeting. I was particularly impressed by the wide variety of speakers and attendees, from tools companies, consultancies, public sector deployers and regulators.

More broadly, since we started the AI Assurance Club in May last year, I’ve been impressed by how seriously those in Europe are taking the safety of AI, whilst still wanting to press ahead with its use. Of course you could argue that this seriousness is being driven by the AI Act, and the severe penalties in the current draft of the AI Liability Directive. However, I think it’s a little wider than that, There is genuine concern and wide concensus in Europe that ‘progress’ should be of benefit to all European citizens, not take place at their expense. One might rail against regulation, arguing that it stifles innovation, makes investors wary and pushes riskier initiatives to more liberal jurisdictions. In some cases that might be so. However if we look at growth for example, we see that over the medium term the EU outperforms the UK, and public services in many EU countries are without doubt superior. There is much greater awareness and consensus around the need for public goods.

I would argue that the legal certainty that results from conformance to standards and regulation in fact drives innovation and adoption of new technology, though I can see an argument that it favours larger players over SMEs for whom this might be a difficult overhead to cover.

More work needs to be done to address that. The risks of AI should not be overstated, but there is clear evidence that  unwanted bias, unwarranted loss of privacy and various other negative outcomes have occurred as a result of unregulated deployment of AI at scale. In my view the EU is leading the way to address these problems, whilst recognising the potential of AI to improve the lives of its citizens. I commend all those in the EU pursuing this in a serious way.