Who is to blame when AI goes rogue?

The New York Times examines who should be held accountable when AI systems behave badly, touching on the thorny question of liability among developers, deployers and regulators.