When an AI system makes a serious mistake, for example it gets into a computer it was never meant to reach, who should hear about it first: the public, a government regulator, or the people using that system?
We don't know the right order. What would you choose?
