Keele University YouTube autism prediction study

Released: July 2019

Can you improve this page?
Share your insights with us

The Atlantic reports that a behavioural study group on children with autism by researchers at Keele University in the UK is but another example of the use of machine learning to 'predict' innate attributes or offer diagnoses.

The study uses AI to study the body movements of children with autism in YouTube videos, using it to classify their behaviours as either typical or atypical, with the aim of more quickly evaluating edge cases that might normally require lab equipment or invasive tactile sensors. 

The Atlantic also points out that the children and parents whose data was scraped by the Keele team had not consented to having their home videos used for scientific research.

The paper was withdrawn 'due to insufficient or definition error(s) in the ethics approval protocol.'

Developer: Andrew Cook; Bappaditya Mandal; Donna Berry; Matthew Johnson; Keele University
Country: UK
Sector: Health
Purpose: Predict autism
Technology: Computer vision; Machine learning; Pattern recognition
Issue: Accuracy/reliability; Privacy; Pseudoscience
Transparency: Privacy