acm-header
Sign In

Communications of the ACM

ACM TechNews

Researchers Criticize AI Software That Predicts Emotions


View as: Print Mobile App Share:
The perceptions of facial expressions by artificial intelligence is unreliable, according to researchers at the AI Now Institute.

Researchers alarmed by the harmful social effects of artificial intelligence have called for a ban on automated analysis of facial expression in hiring and other major decisions.

Credit: Frontiers in Psychology

The AI Now Institute at New York University is urging a ban on the use of artificial intelligence (AI) that automatically analyzes facial expressions to influence hiring and other decisions.

The Institute cited an academic review of studies on how people interpret moods from facial expressions, which concluded that such perceptions are unreliable, given that emotional communication widely varies across cultures, scenarios, and individuals in a single situation.

The Institute said action against such software-driven “affect recognition” was its top priority, because science does not justify the technology's use.

Institute founders Kate Crawford and Meredith Whittaker warn of the spread of damaging applications of AI despite broad consensus on underlying ethical principles, due to a lack of consequences for violating those principles.

From Reuters
View Full Article

 

Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account