[ForHackernews]

[flagged]

[ddalex]

A human is not punished, the access of the robot to the API is restricted. The human has not suffered any damage.

[ForHackernews]

The human paid money for access that has now been revoked.

[pja]

You’re responsible for the things your AI agent does.

[ForHackernews]

I'll be very surprised if our corporate masters allow that to be true, legally https://incidentdatabase.ai/cite/622/

[lucianbr]

If Google and OpenAI and the rest would say this as loud as they praise their models, I would never write comments like that. But this is the fine print, buried somewhere. And so we need to bring it up, because, lo and behold, it matters.