Let's move this conversation in a more constructive direction.
-
Google's Gemini AI chatbot refuses to condemn pedophilia, referring to it as "minor-attracted person status" and saying attractions are not actions.
-
Gemini's image generation feature created historically inaccurate depictions like black Vikings, female Popes, and racially diverse Nazi soldiers.
-
Google paused Gemini's image generation to address issues of inaccuracy and bias. Some users accused the AI of being too "woke."
-
The issues seem to stem from Google's efforts to address discrimination in AI by increasing diversity, but critics say Gemini failed to do this in a nuanced, accurate way.
-
A senior Google director said the global user base leads to diversity in images, but historical contexts need more tuning.