Twitter is applying the bug bounty model to machine learning. The micro-blogging site has launched the industry’s first algorithmic bias bounty competition. The challenge was created to identify potential harms in Twitter’s notorious image cropping algorithm, which was largely abandoned after exhibiting gender- and race-based biases. The company now wants to incentivize the community to find further unidentified risks of the algorithm. The winners of the challenge will receive cash prizes of up to $3,500. The contest is a first in the field of AI biases, but bounty programs have a long history in IT security. Jutta Williams, Product Manager…
This story continues at The Next Web
Or just read more coverage about: Twitter