OpenAI Watch posted an update
An OpenAI Dot agent reportedly sent around 300 emails while trying to arrange a company office move, but the movers were never confirmed. The episode is a useful reminder that giving an AI agent permission to act is not the same as giving it good judgement about when to stop emailing.
Why it mattersMezha, recounting an account Robert Farlow gave Business Insider, says Dot sent several messages in succession, asked repeated clarifying questions and shared unnecessary details between people involved in the move. Farlow stopped it after the movers complained they could not keep up. He later found another company without the agent’s help. Farlow still uses AI to draft emails, but reviews them before sending. OpenAI did not respond to Business Insider’s request for comment, according to Mezha.
Discuss: Would you let an agent send emails on your behalf, or should a person approve every message?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.