Twitter on organization of competition hopes that hackers and researchers will be able to identify the bias in the image cropping algorithm – and will distribute cash prizes to the winning teams (through Engadget). Twitter hopes that by giving teams access to their code and image cropping model, they can find ways in which the algorithm can be harmful (such as cropping in a way that stereotypes or removes the subject of the image).
Competitors want a description of their findings shall be providedand a set of data that can be executed by the algorithm to indicate the problem. Twitter then gives points based on what kind of harm is perceived, how much it can affect people, and more.
The winning team will be rewarded with $ 3,500, and separate $ 1,000 prizes will be awarded for the most innovative and universally qualified discoveries. That amount has caused a bit of a stir On Twitter, a few users say it should be an extra zero. For the context of Twitter a standard bug bounty program would cost you $ 2,940 if you found an error that would allow you to perform functions on someone else (such as Retweeting a tweet or photo) using a cross-site script. Finding an OAuth issue that allows you to take over someone’s Twitter account would get you $ 7,700.
Twitter has conducted its own research on its image cropping algorithm earlier-in May it published a magazine after determining the bias of the algorithm allegations that its previews were racist. Twitter largely deleted crop previews algorithmically after that, but it is still used on the desktop, and a good cropping algorithm is a handy thing for a company like Twitter.
The opening of the competition will allow Twitter to receive feedback from a much broader perspective. For example, the Twitter team held space to discuss the competition, during which a team member mentioned getting questions about caste-based bias in the algorithm, which may not be noticeable to California software developers.
Nor is it just a subconscious algorithmic delusion that Twitter is looking for. The heading has score values for both intentional and unintentional damages. Twitter defines unintentional damage as crops that can result from a “benevolent” user posting a regular image to a platform, while intentional damage is problematic cropping behaviors that someone can take advantage of when posting maliciously designed images.
Twitter says in its information blog that the competition is separate its a bug bounty program – If you submit a report of algorithmic delusions to Twitter outside of the competition, the company will say that your report will be closed and flagged as inappropriate. If you are interested in joining, you can Go to the competition’s HackerOne page to see rules, criteria and more. Applications are open until August 6 at 11:59 p.m. PT, and the winners of the challenge will be announced at Def Con AI Village on August 9.