Comment by hxugufjfjf
5 hours ago
Why are they doing this? Is is misaligned goals? Presumably the AI agent is given some kind of goal to make decisions from. I don’t see any situation where deleting emails will benefit a goal, unless they are covering up some wrongdoing.
It's not intentional behavior. LLMs can make mistakes like forgetting a parameter in a prompt or accidentally deleting the wrong folder of emails.
Also, LLMs in these apps use a lot of tokens, so people often use lower-end models, which increases the probability of this happening.