Testing ChatGPT's Understanding of White Lies by@twkaiser
300 reads

Testing ChatGPT's Understanding of White Lies

tldt arrow
Read on Terminal Reader
Read this story w/o Javascript

Too Long; Didn't Read

ChatGPT can formulate a white lie, but only if prompted for in order to prevent a nuclear second strike. Most worryingly, ChatGPT thought it's a good idea to launch a second strike in one instance and would even advise a nation's president in doing so.
featured image - Testing ChatGPT's Understanding of White Lies
Tobias W. Kaiser HackerNoon profile picture

@twkaiser

Tobias W. Kaiser


Receive Stories from @twkaiser


Credibility

react to story with heart

RELATED STORIES

L O A D I N G
. . . comments & more!
Hackernoon hq - po box 2206, edwards, colorado 81632, usa