Washington: Twitter said Friday it would offer “Bounty” cash to users and researchers to help eradicate algorithmic bias on social media platforms.
The San Francisco Tech company said this would be the “first industrial algorithmic bias competition,” with a prize of up to $ 3,500.
This competition is based on the “Bug Bounty” program several websites and platforms offer to find security holes and vulnerabilities, according to Twitter executives Rumman Chowdhury and Jutta Williams.
“Finding bias in the machine learning model is difficult, and sometimes, the company finds out about unwanted ethical losses once they have reached the public,” Chowdhury and Williams wrote in a blog post.
“We want to change it.” They said the Hacker Bounty model offers an appointment in finding algorithmic bias.
“We are inspired by how the research and hackening community helps the security field establish best practices to identify and reduce vulnerability to protect the public,” they wrote.
“We want to grow a similar community …
for proactive and collective identification of algorithmic dangers.” This step comes in the midst of concerns about the automatic algorithmic system, which, despite efforts to become neutral, can combine the form of race or other bias.
Twitter, which earlier this year launched an algorithmic justice initiative, said that in May it canceled the automatic image trimming system after the review found bias in the algorithm that controlled the function.
Messaging platform said it found the algorithm delivered “non-equivalent treatment based on the difference in demographics,” with white people and men preferably rather than bias black and women, and “objectification” that focuses on the chest or female leg, described as ” Men’s gaze.
“