IBM proposes guidelines to battle AI bias

A set of policy proposals released by IBM on Jan. 21 suggests that government agencies and industry players team up to develop standards to eliminate algorithmic bias in artificial intelligence, Bloomberg reports.

The proposed rules reportedly seek to build a consensus on AI regulation that would land between the stricter standards proposed by government and the more relaxed guidelines favored by technology companies.

Among the recommendations are strategies to ensure algorithms based on historically biased data such as zip codes and mortgage rates do not discriminate against any group, a call for companies to appoint chief AI ethics officers and a request that they keep records of the data fed into algorithms to maintain transparency in AI-led decisions.

"It seems pretty clear to us that government regulation of artificial intelligence is the next frontier in tech policy regulation," Chris Padilla, IBM's vice president of government and regulatory affairs, told Bloomberg, but added that by working alongside the government, tech companies will have a greater influence on the rigidity of that regulation.

"If we take a just-say-no approach or we just wait, the chances are higher that governments will react to something that happens," Mr. Padilla said. "Then you will get more of a prescriptive, top-down regulation."

The proposals came one day before IBM CEO Ginni Rometty is set to lead a panel on AI regulation at the 2020 World Economic Forum in Davos, Switzerland.

More articles on AI:
Viewpoint: Algorithms pitting AI against humans 'make healthcare more dangerous'
Bon Secours Mercy Health invests in behavior change AI platform
Top 10 AI applications for healthcare in 2020: Accenture report

Copyright © 2022 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.

 

Featured Whitepapers

Featured Webinars