AI and cybersecurity (again) --- Guillaume, you mention the use of AI for cyber-attacks, which is becoming critical. There is also the other way round: cyber-attacks which impact the AI decisions - not least about lethal autonomous weapons. Make it even more complex: AI that drives cyber-attacks against AI decisions...

How do we make sure that some critical AI applications are made almost bullet-proof (not 100% secure, of course, but least vulnerable possible)? And that we use AI to combat its misuse for cyber-attacks?

Source author: Vladimir Radunovic vladar@diplomacy.edu