How rogue AI agents became a warning shot about humans losing...

The incidents have revived an old debate in AI safety over keeping an increasingly capable system under human control. But the bigger question is: what does this behaviour actually tell us about an agent’s intentions?