For real, since it’s the company who built the thing and the company who is responsible to control the thing, the company is also responsible when they fail to control the thing. We should stop saying an “agent” did a thing and just say “OpenAI hacks friendly nation in foreign espionage attempt.”
For real, since it’s the company who built the thing and the company who is responsible to control the thing, the company is also responsible when they fail to control the thing. We should stop saying an “agent” did a thing and just say “OpenAI hacks friendly nation in foreign espionage attempt.”
Exactly. That’s all par of anthropomorphizing AI. Don’t blame the agent, blame the people and, most importantly, the company who developed it.