Join getAbstract to access the summary!

Beating Biometric Bias

Join getAbstract to access the summary!

Beating Biometric Bias

The technology is improving – but the bigger issue is how it’s used.

Nature,

5 min read
5 take-aways
Audio & text

What's inside?

Surveillance cameras and facial-recognition systems provoke serious ethical and privacy issues.

  • 0.5×
  • 0.75×
  • 1×
  • 1.25×
  • 1.5×
  • 1.75×
1×

Editorial Rating

9

Qualities

  • Scientific
  • Eye Opening
  • Overview

Recommendation

In TV crime shows, police agents often secretly use fancy facial-recognition software while hiding in black vans and hunting for their suspects. You might wonder whether the police apply facial-recognition software in real life as well. In this eye-opening article in Nature, you will be surprised by how often this relatively new software intervenes in people’s day-to-day lives. The article will engage you if you want to learn more about how this technology works – and think about the difficult ethical questions its application raises.

Summary

Facial-recognition technology is making progress.

Not long ago, digital facial-recognition technology was science fiction. Nowadays it’s deployed by the police to find criminals. Some companies even use it to assess job candidates’ personality traits.

Facial-recognition technologies come in two types. The one-to-one technology makes sure that a person matches their own photo. It’s often used to unlock smartphones or for passport controls. This type of facial-recognition technology has become extremely accurate. Researchers are now also working on implementing 3D facial-recognition techniques as well as skin-texture recognition techniques. In the one-to-many identification, a face is compared with many other faces in a database. This technology is often used by the police to find criminals.

Both technologies have significantly improved thanks to “convolutional neural network” – a kind of deep learning network that is unusually good with images. According to the US National Institute of Standards and Technology, the best algorithms are...

About the Authors

Davide Castelvecchi is a senior reporter at Nature in London. Antoaneta Roussi is a journalist based in Nairobi, Kenya. Richard Van Noorden is a features editor at Nature Research.

Comment on this summary