The case sparked an uproar on social networks. A Facebook recommendation algorithm asked users if they wanted to see other ‘primate videos’ under a UK tabloid video showing black people, the Friday revealed.
New York Times.
“We apologize to anyone who has seen these insulting recommendations,” reacted a spokesperson for Facebook after these revelations, ensuring that the Californian group had deactivated the recommendation tool on this subject “as soon as we noticed this that was happening in order to investigate the causes of the problem and prevent it from happening again ”.
A Facebook recommendation feature powered by AI categorized a video of Black men as “about Primates.” It caused uproar & shame among former & current employees.
Facebook shut off the feature and is investigating. It also apologized after we reached out.https://t.co/p29Bayd5MX
– Ryan Mac 🙃 (@ RMac18) September 3, 2021
“See more videos on primates?” “
The video of Daily Mail, over a year old, is titled “White Man Calls Cops Against Black Men at Marina”. It only shows people, not monkeys. Below, the question “see more videos on primates?” With the options “Yes / Reject” was displayed on the screen of some users, according to a screenshot posted on Twitter by Darci Groves, a former designer of the giant of social networks.
“It’s scandalous,” she commented, calling on her ex-colleagues at Facebook to escalate the matter.
“Our artificial intelligence systems are not perfect”
“Even though we have improved our artificial intelligence systems, we know that they are not perfect and that we have some progress to make,” said the spokesperson for Facebook.
The case highlights the limits of artificial intelligence technologies, regularly highlighted by the platform in its efforts to build a personalized feed to each of its nearly 3 billion monthly users. It also uses it extensively in content moderation, to identify and block problematic messages and images before they are even seen.
But Facebook, like its competitors, is regularly accused of not fighting enough against racism and other forms of hatred and discrimination. The subject arouses all the more tension as many civil society organizations accuse social networks and their algorithms of contributing to the division of American society, in the context of the protests of the movement. Black Lives Matter.