AI companies ignore their own safety rules

Major AI labs are failing to use basic safety measures on the systems they build. Even though these companies talk a lot about safety, they often skip these steps when testing their own work.

This gap between public promises and private action is a big problem. These internal tools are powerful but operate without the oversight that the public expects.

Experts say this needs to change immediately. Companies must hold their own teams to the same standards they set for the rest of the world.