A tester randomly typed “f*** you” into PressArk.
The AI prepared a plan to trash the site content.
It did not execute it, because PressArk forced human approval first.
Funny in testing.
Terrifying in production.
This is why AI agents need real boundaries, approval flows, and a harness that assumes the model can go wrong.
The future is not just “AI can do things.”
The future is “AI can do things safely.”
Try PressArk: pressark.com
abdelali Selouani
Posted on
a "f*** you" prompt caused the agent to try to trash all of the website content !
For further actions, you may consider blocking this person and/or reporting abuse
Top comments (0)