What kind of AI Regulatory Framework will be ideal ?

It will be a bit of both.  However, most of the regulatory framework is already in place.  Much of the mischief is criminal activity that AI makes more effective at scale. The rest is just mischief, like disinformation.  I believe that enforcement by government and business will be a bigger challenge than regulation.  The Twitter file scandal revealed collusion between government and industry to censor political opposition speech.  That goes to show the institutional AI developed to counter AI mischief can be put to ill use as well.  Let the adversarial AI battles begin! 

How are there only two options on this? Shouldn't the framework be more rigorous if human lives are at stake? Regulation for features of self-driving cars should be different than AI computer vision that tells me the rice is almost empty at my favorite burrito joint.