OpenAI’s research on AI models deliberately lying is wild 

Short excerpt below. Click through to read at the original source.

AI models don’t just hallucinate. They also “scheme,” meaning deliberately lie or hide their true intensions.

Read at Source