After a group of researchers previously discovered that the algorithm tends to exclude blacks and men, the Weibo platform Twitter Inc finally revealed its plan to launch a competition for computer researchers and hackers to identify biases in its image cropping algorithm.

This move by Twitter is part of a broader effort by the entire technology industry to ensure that artificial intelligence technologies behave ethically.

The social media platform stated in its blog post that the bounty competition aims to determine “the potential harm of the algorithm is beyond our own recognition.”

Twitter was criticized last year for not including image previews of black faces in its posts. In response, the company said in May that a study by its three machine learning researchers found that an 8% difference in population parity favors women, and 4% favor whites.

Twitter publicly released the computer code that determines how to crop the image in the Twitter feed, and said on Friday that participants were asked to find out how the algorithm caused harm, such as stereotypes or discrediting any group of people.

Winners will receive cash prizes ranging from US$500 to US$3,500 and will be invited to showcase their work at a seminar held on Twitter in August at DEF CON, the largest annual event held in Las Vegas. One of the hacking conferences.

live TV broadcast

#silent