“Going rogue”: Is it time to stop talking about faulty AI frontier models as if they are people?

We should take care not to allow our desire to attach human qualities to the non-human to mask the serious issue at hand—who is accountable when AI models go wrong?