'Fear' - Amazon facial recognition can detect a new emotion in our eyes
Amazon says that its facial recognition technology can now analyse a new emotion: fear.
The tech giant, whose ‘Rekognition’ technology is licensed for use for a variety of purposes including security and policing, announced the update this week in a blog post.
At present, the face identification software can spot seven emotions: ‘happy’, ‘sad’, ‘angry’, ‘surprised’, ‘disgusted’, ‘calm’ and ‘confused’.
“We have improved accuracy for emotion detection and added a new emotion, fear,” said Amazon in a written update.
“Face analysis generates metadata about detected faces in the form of gender, age range, emotions, attributes such as ‘Smile’, face pose, face image quality and face landmarks.”
Facial recognition technology is of interest to a number of different industries, from retail and tech to security.
But the move to add ‘fear’ to the list of emotions catalogued will add to unease among civil liberty groups about the growing sophistication of facial recognition and its potential misapplication.
In the past year Amazon has found itself at the centre of a growing debate over the use of facial recognition by governments, with critics warning of false matches and arrests and proponents arguing it keeps the public safe. Police in some US states have used Amazon’s face and image ID service, known as Rekognition.
Amazon is one of many facial recognition vendors, alongside France's Idemia, Japan's NEC and Microsoft. In May, Amazon shareholders rejected proposals to curb and audit its facial recognition technology, particularly in offering it to governments unless the company’s board determined that such sales did not violate civil liberties.
Privacy campaigners have been warning about the danger of storing facial recognition records.
Elsewhere this week, a biometric database used by the UK Metroplitan police was exposed, leaving facial recognition and fingerprint data of over a million people vulnerable to misuse.
The database vulnerability, associated with the web-based Biostar 2 biometrics lock system typically used by businesses and warehouses, was discovered by two Israeli researchers who gained access 28m records including fingerprints and facial recognition data.