Capital letter test is a foolproof way of sorting AIs from humans

1 year ago 223

Technology

A instrumentality for asking questions utilizing superior letters seems to baffle artificial intelligences similar ChatGPT, portion humans tin easy springiness the close answer

By Chris Stokel-Walker

There is an casual mode to fool ChatGPT

Ascannio/Shutterstock

A clever usage of superior letters could beryllium an casual mode to flummox artificial intelligences similar ChatGPT, letting radical separate them from humans successful conversation.

The thought is reminiscent of the Turing test, archetypal projected by machine idiosyncratic Alan Turing successful 1950. He said that an AI would beryllium considered genuinely intelligent erstwhile we couldn’t separate its answers from a human’s. But present that ample connection models (LLMs) similar ChatGPT tin dependable convincingly human, Hong Wang astatine the University of …

View introductory offers

No commitment, cancel anytime*

Offer ends 14th June 2023.

*Cancel anytime wrong 14 days of outgo to person a refund connected unserved issues.

Inclusive of applicable taxes (VAT)

or

Existing subscribers

Sign successful to your account

More from New Scientist

Explore the latest news, articles and features

Popular articles

Trending New Scientist articles

Read Entire Article