This is similar to what happened with Tay.
I don’t remember Tay, even though it was less than a decade ago, but I do remember people abusing AI a couple of years ago by getting it to render images of black people wearing Nazi uniforms, and other chicanery.
FTA
Media analyst Mark Dice exposes how the Grok chatbot has become so paranoid that it’s
now seeing “Nazi ideology” in photos of vegetables and racist tropes in pictures of alligators.
The responses are so absurd that Grok makes the ADL’s Jew
worshipper, Jonathan Greenblat, look reasonable by comparison.