>>vforgi+(OP)
I understand the idea behind it: the risks are high and we want to ensure that the AI can not be used for purposes that threatens the survival human civilization. Unfortunately, there is high probability that this agency will be abused from day one: instead of (or in addition to) focusing on the humanity's survival, the agency could be used as a thought police. The AI that allows for the 'wrongthink' will be banned. Only the 'correct think' AI will be licensed to the public.