facebook has apologized in a statement for its AI’s “unacceptable flaw” as its black men are mislabeled as “primate” content. It has also disabled the recommendation feature that was responsible for the message as it investigates the root cause to prevent serious errors like this from happening again.
Facebook’s AI put a horrific label on a video of black men by asking users who had watched a Daily Mail video featuring black men a quick question if they wanted to “keep watching videos about primates.”
Company spokesperson Dani Lever said in a statement: “As we have said, we know that while we have made improvements to our AI, it is not perfect and we need to make even more progress. We apologize to anyone who has made these offensive recommendations. has seen. “
Gender and racial bias in artificial intelligence is hardly a problem unique to the social network – facial recognition technologies are still far from perfect and tend to misidentify POCs and women in general. Last year, fake facial recognition contests led to the wrongful arrests of two black men in Detroit. In 2015, Google Photos tagged the photos of black people as “gorillas,” and Wired discovered a few years later that the tech giant’s solution was to censor the word “gorilla” from searches and image tags.
A few months ago, Facebook shared a dataset created with the AI community in an effort to combat the problem. It featured over 40,000 videos with 3,000 paid actors sharing their age and gender with the company.
Facebook has even hired professionals to light their shoots and label their skin tones so AI systems can learn what people of different ethnicities look like under different lighting conditions. The recent event made it clear that Facebook has not fully resolved the AI bias.