• | 4:00 pm

This company’s software is designed to protect against AI disasters

Robust Intelligence stress-tests machine learning models and identifies bad data.

This company’s software is designed to protect against AI disasters
[Source photo: Getty Images]

As machine learning transforms business processes across an array of industries—from finance to medicine to travel—it’s essential to root out flaws which, if left unchecked, could have catastrophic consequences. Robust Intelligence is a Bay Area company focused on mitigating AI risks, helping organizations continuously validate AI models and data, and looking for anomalous behaviors and weaknesses. This allows data science teams to more quickly and efficiently test the robustness of a model’s performance, fairness, and security. Robust Intelligence’s AI Firewall software can also prevent bad data from entering production models by looking for anomalies in real time.

Last year, Robust Intelligence won new customers that include JPMorgan Chase, Deloitte, and Hitachi; cosponsored a pair of competitions focused on algorithmic integrity; and helped back a first-of-its-kind bug bounty focused on AI bias. The competition was organized by a group of volunteers who work in Big Tech, part of a growing consciousness that cofounder and CEO Yaron Singer applauds as fundamental to driving change. “We want everybody to say, ‘How can we protect users who are affected by these models, and make sure that we’re delivering this in the safest way possible?’ ”

  Be in the Know. Subscribe to our Newsletters.


Alex is a contributing editor at Fast Company, the founding editor and editor at large of Motherboard at Vice, and a freelance writer and producer with a focus on the intersections of science, technology, media, politics, and culture. More

More Top Stories: