Amazon Rekognition Learns to See Human Fear

Amazon is claiming that it can see fear thanks to improvements to its Rekognition facial recognition platform. The latest update introduces more accurate gender and age identification. It will also be better at reading people’s emotions based on their facial characteristics, adding fear to a list that already includes happy, sad, angry, surprised, disgusted, calm, and confused.

Rekognition is available as part of the Amazon Web Services suite, although the system has come under fire in recent months due to concerns about public privacy. Amazon has repeatedly defended the sale of surveillance tech to police departments despite vocal criticism from the ACLU as well as Amazon employees and shareholders alike.

The latest improvements are unlikely to alleviate those critiques, which emphasize the ethics of facial recognition rather than its accuracy. It’s also worth noting that Amazon is recovering from a rather high-profile failure after the Orlando Police Department abandoned its trial of a Rekognition surveillance program, largely because they were never able to get the system up and running or perform a live test.

Rekognition can be used to analyze people’s faces in videos and still images. Police departments may be interested in a more accurate facial recognition platform, although Amazon itself did not specify any intended use cases when announcing the updates.

Sources: The Next Web, CNBC

–

August 14, 2019 – by Eric Weiss