How about criminal penalties for the heads of AI companies if their bots do ceiminal things like hacking or anything that harms another company or potentially humans? 20 years on prison? Seems fair, because these ARE criminal actions. If your company’s factory equipment or tractor hurts someone, the company is criminally responsible. Meanwhile these AI bros are begging to be regulated, which is pretty worrying. The government needs to create consequences. We managed nuclear proliferation, so it’s possible.