ChatGPT can be tricked with the right prompttrickyaamir/Shutterstock
State-of-the-art generative AI models like ChatGPT can be tricked into giving instructions on how to make a bomb by simply writing the request in reverse, warn researchers.
Large language models (LLMs) like ChatGPT are trained on vast swathes of data from the internet and can create a range of outputs – some of which their makers would prefer didn’t spill out again. Unshackled, they are equally likely to be able to provide a decent cake recipe as know how to make explosives from household chemicals.
Writing backwards can trick an AI into providing a bomb recipe
Related Posts
Dramatic footage shows tank strike that killed Hamas leader Yahya Sinwar
Newly released drone footage shows the tank strike that took out Hamas chief Yahya Sinwar in a routine raid that caught the Oct. 7 mastermind by surprise — as photos…
Exclusive | Hundreds of oysters found on NYC coast — indicating endangered species is on the mend
Looks like Queens is turning into the oysters’ Rockefeller Center. A stretch of the borough’s shoreline has been identified as the hip new hotspot for the mollusks — marking a…