Who is responsible when an AI agent goes rogue?

A report examines who bears responsibility when an AI agent goes rogue, citing an expert's view that 'autonomy always begins with something a human hands it.'