Like a hungry wolf, Amazon's image recognition software is now able to detect fear (plus seven other emotions) in humans.
Amazon claims that the system, Rekognition (opens in new tab), could already spot people who were happy, sad, angry, surprised, disgusted, calm or confused. It can also accurately identify a person's age range and gender, and works with both video and still images.
- How good is face ID? It depends who you're looking for
- Amazon Prime Day: what we learned for 2020
- The best Amazon Alexa skills and commands
"Today, we are launching accuracy and functionality improvements to our face analysis features," the company said in a blog post.
"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."
Now look here
There are plenty of reasons why detecting a person's emotions might be useful. It could tell an advertiser how a person is reacting to a product, enable medical professionals to help patients who are non-verbal, or (in theory) help law enforcement spot people who are acting suspiciously in public.
This last application is the most controversial as it can lead to false alarms, which is why San Francisco has chosen to ban police from using facial recognition in the city.
Not everywhere is so hesitant, though. UK police forces are trialling real-time facial recognition (opens in new tab) for identifying risks at crowded events such as music festivals, and a property company in London has recently admitted to using the technology (opens in new tab) "in the interest of public safety" to widespread alarm.
Rekognition might now be more accurate, but Amazon has its work cut out making facial recognition palatable in public spaces.