‘White’ artificial intelligence risks exacerbating racial inequalities, study finds | Scientific and technical news


The “whiteness” of artificial intelligence (AI) risks creating a “racially homogeneous” workforce as humans create machines warped by their biases, a study suggests.

The Cambridge University study looked at AI in society, including in movies, Google searches, stock footage, and robot voices.

The researchers have suggested that the machines have distinct racial identities, which perpetuates racial stereotypes of the “real world”.

Flash Gordon’s Ming the Merciless is an example of “extraterritorial racial stereotypes,” according to the study.

According to the researchers, non-abstract AI in internet search engine results typically had Caucasian characteristics or was white in color.

Most of the virtual voices in the devices spoke in “standard white middle class English” because “ideas of adding black dialects were dismissed as being too controversial or outside the target market,” the study concludes.

Experts analyzed recent research in a variety of fields, including human-machine interaction and critical race theory, to demonstrate that machines can be “racialized” and that this perpetuates “real-world” racial biases.

This includes work on how bots are perceived to have distinct racial identities, with black bots being more abused online, and research showing that people feel closer to virtual agents when they perceive a shared racial identity.

Learn more about artificial intelligence

According to scientists at Cambridge’s Leverhulme Center for the Future of Intelligence (CFI), like other science fiction tropes, artificial intelligence has long reflected racial thinking.

They pointed to extraterritorial racial stereotypes such as Ming’s ruthless “oriental” alien to the Caribbean caricature of Jar Jar Binks.

However, they suggested that AI is depicted as white, because unlike species on other planets, it has attributes used to “justify colonialism and segregation” in the past.

Co-author Dr Stephen Cave said the most powerful on-screen robots were either white or played by white actors, including Terminator, Blade Runner, Metropolis, and Ex Machina.

“Metal and plastic androids have white lines, like in I, Robot. Even disembodied AIs – from HAL 9000 to Samantha en Her – have white voices, ”he said.

“It wasn’t until very recently that a few TV shows, such as Westworld, used AI characters with a mix of skin tones. ”

:: Listen to the daily podcast on Apple Podcasts, Google Podcasts, Spotify, Spreaker

Principal investigator Dr Kanta Dihal said: “As society has, for centuries, promoted the association of intelligence with white Europeans, one should expect that when asked this culture of imagining an intelligent machine, it imagines a white machine.

“People trust AI to make decisions. Cultural representations encourage the idea that AI is less fallible than humans. ”

She warned that if the demographics of AI did not diversify, racial inequalities would make it harder for non-whites to advance in tech.

“In cases where these systems are racialized as white, it could have dangerous consequences for humans who are not,” she said.

Dr Dihal concluded: “The perceived whiteness of AI will make it more difficult for people of color to move forward on the pitch.

“If the developer demographics do not diversify, artificial intelligence risks exacerbating racial inequalities. “


Please enter your comment!
Please enter your name here