China facial image criminality inference

November 2016
Updated: October 2021

Researchers at Shanghai Jiao Tong University have released the results (pdf) of research that suggests that software can automatically detect with 85.5% certain that a human is a criminal by analysing their facial features.

The results, which indicate that people with smaller mouths, curvier upper lips and closer-set eyes are more likely to be criminals, has resulted in accusations of inaccuracy and bias, and the researchers of conducting phrenology and pseudoscience.

In 2020, a similar study by researchers at Harrisburg University had to be withdrawn after an academic and industry backlash.

Developer: Xiaolin Wu, Xi Zhang
Purpose: Recognise/predict criminality
Technology: Facial analysis
Accuracy/reliability; Bias/discrimination - race, gender, age, income; Pseudoscience