Harrisburg University study predicts criminality based on one facial picture

Released: June 2020

Can you improve this page?
Share your insights with us

Professors and a PhD student at Harrisburg University developed software that automatically predicted whether someone would become a criminal based solely on a picture of their face with '80 percent accuracy and no racial bias'.

A backlash quickly followed the announcement, with the researchers accused of 'unsound scientific premises, research, and methods which … have [been] debunked over the years' by the Coalition for Critical Technology (CCT) in an open letter signed by over 1,700 academics demanding the research remain unpublished.

The research was intended to appear in a book series titled 'Springer Nature – Research Book Series: Transactions on Computational Science & Computational Intelligence.' 

Springer Nature later confirmed it would not publish the research, which was subsequently withdrawn by Harrisburg University.

Developer: Harrisburg University
Country: USA
Sector: Govt - police
Purpose: Predict criminality
Technology: Facial recognition; Emotion detection
Issue: Accuracy/reliability; Bias/discrimination - race, gender, age, income; Ethics

Page info
Type: Research
Published: January 2023
Last updated: February 2024