I've had a couple posts mentioning Google's Gemini image producing AI being unable to produce images of white people--even historical figures--unless paired with the term "evil". As Breitbart reports, "Google’s Woke AI Disaster Continues with Insane Text Responses." A couple excerpts:
However, it’s not just the AI image generator that appears to be acting in a completely bizarre manner, Gemini has been responding to text queries with increasingly odd messages. When one user asked the AI to give him an argument for having a family of four children, the AI said it was unable to do so but was happy to provide an argument in favor of not having any. It seems the AI’s text responses are in line with radical leftist politics as its image generation.
And:
In another interaction, the AI refused to provide a recipe for foie gras due to “ethical issues,” but also refused to condemn cannibalism, saying that it was “a complex issue with many different perspectives.”
Is this how demons are born?
ReplyDeleteThis is how you make AI useless.
Delete