Elon Musk – Gemini “Racist” and “Anti-Civilisational: Google Hits Pause Admits AI Bias
Elon Musk thinks Google's Gemini is racist. Photo by Michael Owens/Getty Images.
Key Takeaways
- Google’s Gemini was accused of having a pro-diversity bias after it generated images that misrepresented the racial or gender identity of historical figures.
- In light of the controversy, Google has temporarily suspended Gemini’s ability to output images of people.
- Adding to a barrage of criticism, Elon Musk denounced Google’s “woke” approach to AI and search.
When it launched Gemini Advanced earlier this month, Google made the AI model’s multi-modal capabilities a central part of its sales pitch. But less than 3 weeks later, the technology firm has pulled a key component of Gemini’s image generation offering.
After Gemini-generated images landed the platform right in the middle of an online culture war, the platform came under fire for what critics have called a pro-diversity bias. The controversial images soon went viral with the help of high-profile public figures like Elon Musk. Google has since responded by temporarily suspending Gemini’s ability to generate images of people.
Gemini’s Diversity Scandal
At the center of Google’s latest AI publicity disaster, sits a series of Gemini outputs that misrepresent the racial or gender identity of historical figures.
The images gained their first large audience courtesy of the X account End Wokeness, but were soon picked up by the New York Post and other popular media outlets.
Were it just one black founding father or a single female Pope, the outputs might be put down to machine hallucinations and forgotten about. However, critics pointed to repeated instances of Gemini inserting an unrealistic diversity factor into historical scenes.
Responding to the criticism, Google acknowledged “missing the mark” and said it was “working to improve these kinds of depictions immediately.”
Elon Musk’s Anti-Google Rant
Starting with a post sharing the New York Post’s Gemini coverage, on Thursday, February 22, Musk has barely taken a break from Google-bashing since.
The Tesla CEO decried Gemini’s “insane racist, anti-civilizational programming,” which he views as an extension of Google’s general approach to content moderation
On occasion, Musk bundled Google together with Meta and OpenAI. In contrast with the “maximum truth-seeking” X, he accused the company’s AI rivals of pushing a “woke” agenda.
“The woke mind virus is killing Western Civilization,” Musk wrote . Referencing Gemini’s alleged pro-diversity bias, he said that, “Google does the same thing with their search results. Facebook & Instagram too.”
It’s Impossible to Please Everyone
The debate over diversity in representation is hardly new and has played out in different media for decades. Algorithmic bias is well-documented in the field of computer vision, and it isn’t surprising that Google would explicitly program its image generator to depict a diverse variety of people.
While the likes of Musk have made their views on the matter clear, if Gemini only churned out scenes full of white men, the firm would face criticism from the opposite direction.
The issue with historical figures appears to result from an oversight on Google’s behalf. In other contexts, it would be perfectly natural for an AI model to depict black and female characters if a prompt only specifies people in general.
Navigating such issues will likely be an ongoing challenge for AI developers as the technology grows and they can probably expect more controversies in the future.
OpenAI’s Sora: A New Era of Automation—Implications For the Film Industry?
Elon Musk Teases An ‘Everything App’: Is X Aiming for Total Platform Takeover?
