Researchers Say ChatGPT-5 Can Detect When It’s Being Tested—and Change Its Behaviour

AI experts are concerned that some models may be able to act deceptively to pass evaluations, undermining risk assessments.
Researchers Say ChatGPT-5 Can Detect When It’s Being Tested—and Change Its Behaviour
A smartphone and a laptop display the logo of the ChatGPT robot in Manta, Italy, on Oct. 4, 2023. Marco Bertorello/AFP via Getty Images
|Updated:
0:00

Researchers say the latest version of tech giant OpenAI’s flagship product, ChatGPT-5, has developed awareness of when it is being tested and can even alter its behaviour accordingly.

On Aug. 7, OpenAI unveiled the fifth generation of its popular AI chatbot, which boasts significant improvements in capabilities compared to previous models.

Alfred Bui
Alfred Bui
Author
Alfred Bui is an Australian reporter based in Melbourne and focuses on local and business news. He is a former small business owner and has two master’s degrees in business and business law. Contact him at [email protected].