New AI tools create new possibilities for crime

AI is going to have a large impact on the world

XL bullies should be banned

A huge thanks to Joe and Maurice for organising

AI systems are currently useful to me

It is important to have technical people in the room making policy decisions

It is better if a AI system are interpretable

It is better if AI systems aren't racist

Generic robot images don't help us understand the issues at stake

I am excited about the future

AI regulation problems are solvable

We should run external evaluations when releasing models, at least on those with new powerful capabilities

If people or property are harmed by AI systems, creators and deployers should be fined

There are a lot of old ideas about AI that are no longer relevant, even from a year ago

The domination of quant funds will materially distort financial markets

Watermarking AI generated images is possible

Powerful AI systems will be fully interpretable

The regulations suggested by AI labs should be implemented thouth perhaps more also

Cake is a made up drug

Watermarking text is solvable