‘Horrible experiment’ shows that Twitter’s cropping algorithm is racially biased

Social media giant Twitter on Monday said that it would investigate its image cropping algorithm after many users complained that it favoured white faces over black. The tool crops pictures that are too big to fit on the screen – and automatically selects which parts of an image should be cut off.

The ‘racist’ nature of the algorithm was first noticed by PhD student Colin Madland. Madland, who is White noticed that Twitter automatically cropped his Black colleague whenever he posted a screenshot of a zoom call.

Also Read | Twitter expands QAnon restrictions to political candidates, elected officials

Subsequently, the experiment was repeated by entrepreneur Tony Arcieri, using a collage of US senator Mitch McConnell and Barack Obama. And he found, the algorithm consistently picked McConnel’s image.

“Twitter is just one example of racism manifesting in machine learning algorithms,” Arcieri said after the experiment.

Twitter which noticed the series of subsequent tweets that corroborated the claim said in a statement that their team had tested for bias before shipping the model and did not find evidence of racial or gender bias, However, acknowledging the mistake it further vowed to investigate the issue.

Also Read | These emojis were the most popular during quarantine, reveals Twitter

“It’s clear that we’ve got more analysis to do. We’ll continue to share what we learn, what actions we take and will open-source it so others can review and replicate,” they said.

Twitter is not the first social media platform that has been called out for racism. Earlier in 2015, Google was criticised for an image-recognition algorithm that auto-tagged pictures of black people as “gorillas.”