Humans Must Have Operational Control

Who is responsible for that action an AI agent takes? Well it is you and me. Someone is responsible. In the same way a gun does not shoot someone of its own volition if we release automation that no one has operational control over. If we release systems that can draw conclusions given data they will be wrong.

If you automated a weapon that used facial recognition and was only supposed to target individuals in a database, then it shoots a child because they look vaugly similar well then we are responsible. That buck that everyone passes is going to stop somewhere.

Then comes the mob.

This line appears after every note.