[ad_1]
\n<\/p>\n
Meta AI is consistently unable to generate accurate images for seemingly simple prompts like \u201cAsian man and Caucasian friend,\u201d or \u201cAsian man and white wife,\u201d The Verge<\/em> <\/a>. Instead, the company\u2019s image generator seems to be biased toward creating images of people of the same race, even when explicitly prompted otherwise.<\/p>\n Engadget confirmed these results in our own testing of Meta\u2019s <\/a> image generator. Prompts for \u201can Asian man with a white woman friend\u201d or \u201can Asian man with a white wife\u201d generated images of Asian couples. When asked for \u201ca diverse group of people,\u201d Meta AI generated a grid of nine white faces and one person of color. There were a couple occasions when it created a single result that reflected the prompt, but in most cases it failed to accurately depict the prompt.<\/p>\n As The Verge<\/em> points out, there are other more \u201csubtle\u201d signs of bias in Meta AI, like a tendency to make Asian men appear older while Asian women appeared younger. The image generator also sometimes added \u201cculturally specific attire\u201d even when that wasn\u2019t part of the prompt.<\/p>\n It\u2019s not clear why Meta AI is struggling with these types of prompts, though it\u2019s not the first generative AI platform to come under scrutiny for its depiction of race. Google\u2019s Gemini image generator paused its ability to create images of people after it overcorrected for diversity with <\/a> in response prompts about historical figures. Google <\/a> that its internal safeguards failed to account for situations when diverse results were inappropriate.<\/p>\n Meta didn\u2019t immediately respond to a request for comment. The company has previously described Meta AI as being in \u201cbeta\u201d and thus prone to making mistakes. Meta AI has also struggled to accurately answer <\/a> about current events and public figures.<\/p>\n<\/div>\n