IBM launches real-time algorithmic bias detection tools

Sep 19, 2018 | Chris Burt

Categories Biometric R&D | Biometrics News

IBM launches real-time algorithmic bias detection tools

IBM has launched software to analyze how and why algorithms make decisions, as well as detect bias and recommend changes. The new Trust and Transparency in AI service is a set of capabilities served from the IBM Cloud, while the Fairness 360 Kit in an open source library of algorithms, code, and tutorials designed for use by academics, researchers, and data scientists.

Recent research from IBM indicates that 82 percent of enterprises are considering deploying AI, but 60 percent are concerned about liability, and 63 percent do not have the necessary in-house talent to manage the technology.

The Trust and Transparency capabilities operate as a real-time automated service, which detects potentially unfair outcomes and recommends data to add to the model to mitigate any bias. It also provides a record of model accuracy, performance, fairness, and lineage, which IBM says may be helpful for GDPR compliance.

“IBM led the industry in establishing Trust and Transparency principles for the development of new AI technologies,” said Beth Smith, General Manager of Watson AI at IBM. “It’s time to translate principles into practice. We are giving new transparency and control to the businesses who use AI and face the most potential risk from any flawed decision making.”

By releasing AI Fairness 360 as a companion open source project to the Adversarial Robustness Toolbox, IBM hopes to encourage the contributions of researchers outside the organization, and eventually build a community dedicated to improving algorithms, according to a blog post.

IBM also announced plans earlier this year to launch a massive data set to improve facial recognition and reduce bias, and other tech giants including Google and Microsoft have released tools for developers.

Independent efforts to improve algorithm fairness include the Algorithmic Justice League, which was launched by MIT researcher Joy Boulamwini, who’s report on facial recognition algorithms from Microsoft, IBM, and Megvii sparked the controversy earlier this year.