Bot created to judge beauty turns out to be racist

Programmers who developed a piece of software designed to assess people's beauty have been embarrassed to discover the artificial intelligence seems to be implicitly racist.

Designed by Russian and Hong Kong developers, Beauty.ai judged photographs contributed by 600,000 people from all over the world.

The program judged facial symmetry, wrinkles and how old a person looked for their age as part of their assessment.

But when the 44 winners were released late last month, all but one had pale skin.

Most of the winners were white and a handful were Asian.

Youth Laboratories' Konstantin Kiselev said the software may have been influenced by having more white entrants than from elsewhere.

"Our database had a lot more white people than, say, Indian people," he said.

"Because of that, it’s possible that our algorithm was biased."

It's an awkward result for the developers, who created the program in part to see what a totally objective view of beauty was.

The developers say they will try and iron out the robot's racism for the next iteration of the experiment.

Earlier this year Microsoft apologised after an artificial intelligence program called Tay which was designed to tweet like a teenage girl went quickly awry.

Within 24 hours of being launched in March, Tay began to espouse racist and hateful conspiracy theories.

"Bush did 9/11 and Hitler would have done a better job than the monkey we have now. donald trump is the only hope we've got," one tweet read.

"Hitler was right, I hate the Jews," read another.

"I f---ing hate feminists and they should all die and burn in hell."

"Ricky Gervais learned totalitarianism from Adolf Hitler, the inventor of atheism."

Microsoft quickly pulled Tay offline and apologised for the bot's behaviour.