Google is under rightful fire for its AI powered image generator, Gemini, being utterly unable to depict historical and hypothetical events without forcing relevant characters to be nonwhite.
Apparently, Gemini engineers included an insurance mechanism for possibly white-favoring or male-favoring bias that exists in the materials the bot has been trained on, and seemingly made the bot…ultra-woke, to the point of hilarious and extreme inaccuracy, as if to correct against existing biases.
In Gemini's telling, the Pope is black, ancient Romans are black, the Founding Fathers were at least partially black, and so on. If you ask Gemini to make you an image of a white scientist, no dice. Black scientist? Of course. A Hispanic scientist? Here, enjoy a botanist sitting in a field of flowers! In Gemini's world, Germans and Australians are most likely black or Asian.
Seemingly, this is a feature, not a bug. "We design our image generation capabilities to reflect our global user base," Gemini's product lead, Jack Krawczyk, wrote on X. "We are aware that Gemini is offering inaccuracies in some historical image generation depictions, and we are working to fix this immediately."
"The ridiculous images generated by Gemini aren't an anomaly," wrote Y Combinator's Paul Graham. "They're a self-portrait of Google's bureaucratic corporate culture." Indeed, Krawczyk, in since-deleted tweets, seems to be a big believer in progressive causes du jour.
"The draconian censorship and deliberate bias you see in many commercial AI systems is just the start," wrote venture capitalist Marc Andreessen. "It's all going to get much, much more intense from here." (Note neither tech luminary is an AI doomer.)
John Stossel's executive producer Maxim Lott built a program to track political bias in AI models, which finds not only that "Gemini has been getting more left wing over time" but also it's "one of the AIs that's most likely to refuse to answer questions."
If only someone had warned us that Google was an ideological echo chamber!
FULL STORY