The AI image generator developed by Meta, known as Imagine, is facing accusations of racial bias. Users have discovered that the tool is unable to produce pictures depicting an Asian man with a white woman. This bias is surprising, considering that Mark Zuckerberg, Meta's CEO, is married to a woman of East Asian heritage. The AI-powered image generator, Imagine, was launched in late 2023 and has the capability to transform written prompts into realistic pictures almost instantly. However, users found that it consistently failed to generate images featuring mixed-race couples. When prompted for an image of an Asian man with a white wife, the tool only displayed pictures of Asian couples. Some users even shared pictures of Zuckerberg and his wife, joking that they were able to achieve the desired image manually. The Verge first reported this issue, with reporter Mia Sato stating that she attempted to create images of Asian men and women with white partners or friends dozens of times, but only received accurate results once.

Meta has not yet responded to the request for comment from Business Insider regarding this matter. It's worth noting that Meta is not the first tech company to face criticism for "racist" AI. Google had to temporarily halt its Gemini image generator when it was discovered that the generated images were historically inaccurate, including depictions of Asian Nazis in 1940 Germany, black Vikings, and female medieval knights. These incidents highlight the ongoing concerns regarding racial biases in AI algorithms. When predictive algorithms heavily rely on crowdsourced opinions, stereotypes, or biased data, they can perpetuate exclusion, stereotypes, and unfair targeting of individuals and communities. Generative AIs like Gemini and Imagine draw from large amounts of training data sourced from society. If there is a lack of representation of mixed-race couples in this data, it could explain the AI's struggles in generating those types of images.