Sign In

Communications of the ACM

ACM TechNews

How to Design AI that Eliminates Disability Bias


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
Eliminating AI Bias

Tackling the multi-faceted needs of employees requires new learning systems.

Credit: Financial Times

As artificial intelligence (AI) is added into devices and services, the technology has proven to magnify the prejudices that skew human decision making against women and ethnic minorities.

In addition, AI has been shown to have negative repercussions for those with disabilities. The inability of AI to handle the unexpected threatens to worsen the historic underemployment of disabled people by robbing candidates of job opportunities, according to a recent report by the AI Now Institute at New York University.

The report highlights the popularity of remote video interviewing technologies, which use AI to analyze speech patterns, facial expressions, and candidates' tone of voice to draw inferences about their employability. Critics note these systems risk disadvantaging applicants with disabilities, such as speech disorders, facial paralysis, or autism.

From Financial Times 
View Full Article - May Require Paid Subscription

 

Abstracts Copyright © 2020 SmithBucklin, Washington, DC, USA


 

No entries found