/cdn.vox-cdn.com/uploads/chorus_asset/file/23985505/acastro_STK024_01.jpg "Major Reddit Communities Will Shut Down To Protest Threat To Third-Party Apps")
There is an easy way to trick ChatGPT
Ascannio/Shutterstock
Clever use of capital letters could be an easy way to baffle artificial intelligences like ChatGPT, allowing people to distinguish them from humans in conversation.
The idea is reminiscent of the Turing test, first proposed by computer scientist Alan Turing in 1950. He said that an AI would be considered truly intelligent once we couldn’t distinguish its responses from those of a human. But now that extended language models (LLMs) like ChatGPT can sound convincingly human, Hong Wang of the University of…
