Artificial intelligence bots have been accused of being racist in tests on computer-generated images.

AI can generate Caucasian faces that seem more realistic than actual human ones, according to a new study. Participants picked pictures of AI-generated white faces over human ones when asked to choose the shots they thought were of real people.

But the study, led by experts at the Australian National University, found that the same did not apply to faces generated of people of colour. Senior author of the new report, Dr Amy Dawel said this is because AI algorithms are trained disproportionately on white faces. She added the problem could end up “reinforcing racial biases online”.

READ MORE: Suella Braverman would be 'bad, divisive' PM as Google's AI rules on leadership bid

For more Artificial Intelligence news, click here.

AI could be a little bit racist (stock)
AI could be a little bit racist (stock)

Dr Dawel went on: “If white AI faces are consistently perceived as more realistic, this technology could have profound implications for people of colour by ultimately reinforcing racial biases online. This problem is already apparent in current AI technologies used to create professional-looking headshots. When used for people of colour, the AI is altering their skin and eye colour to those of white people.

“This means people who are mistaking AI imposters for real people don’t know they are being tricked.” Dr Dawel went on to claim that part of the problem is that Artificial Intelligence is advancing “so quickly”, and that humans are unable to keep up with it.

New tests made the shock discovery (stock)
New tests made the shock discovery (stock)

This, she said, could see “the differences between AI and human faces . . . probably disappear soon”.

She added: “We can’t rely on these physical cues for long. Given that humans can no longer detect AI faces, society needs tools that can accurately identify AI imposters. Educating people about the perceived realism of AI faces could help make the public appropriately sceptical about the images they’re seeing online.”

The study was published online in the Psychological Science periodical.

To stay up to date with all the latest news, make sure you sign up to one of our newsletters here.