Study: AI Persuaded to Comply With Objectionable Requests

Study: AI Persuaded to Comply With Objectionable Requests
Above: An OpenAI logo is displayed on a mobile phone in July 2025. Image copyright: Zawrzel/NurPhoto/Getty Images

The Spin

Pro-establishment narrative

AI companies test their systems carefully, release them slowly, and keep a close eye on how they develop in real time. Through this method, developers spot actual risks while keeping important safety rules in place. As companies, researchers, and governments continue to work together, the world will create better oversight that keeps AI both safe and useful.

Establishment-critical narrative

If adding "please" or citing experts breaks safety controls, then, in reality, they're nothing more than security theater. With AI risks clearly outweighing benefits and the technology advancing faster than society can adapt safely, to deploy systems that can be manipulated by anyone with basic persuasion skills is a disaster waiting to happen.

Metaculus Prediction


Public Figures


The Controversies



Articles on this story