Comment by f6v

I think that's a reasonable concern. At the same time, I think there will be many careless people who will give the agents access to everything. Some of them will suffer when an agent fails. But that's more training data for OpenAI.