AI is Reducing the Cost of Biometric Authentication

Developments in Artificial Intelligence are helping to reduce the cost of biometric identification and authentication techniques.

Historically voice scans, fingerprint scans, face scans and retina scans have either been too expensive to implement on a large scale, or too susceptible to manipulation.

Early face recognition tech, for example, could be fooled by a printed photo of someone’s face.

There were further issues with functionality reducing under poor lighting conditions, and some high-profile controversies around facial recognition and ethnicity.

AI is able to improve the effectiveness of such technology. Apple’s new iPhone X, for example, uses infrared to build a complex, three-dimensional model of a user’s face.

The Next Web reports that modern iterations can now adapt to changing facial features (such as facial hair), and detect whether or not a user is sleeping.

The improvements have come with the gradual integration of neural networks. The iPhone X comes with a neural network processor built in.

More nuanced kinds of authentication are also becoming viable – AI is able to help detect atypical typing and clicking patterns, browsing habits and user behaviour within a certain account, and ask for additional verification steps if activity deviates from the norm.

The Next Web argues: “The addition of AI behavioral analytics smarts to online account protection will make it exponentially more difficult for bad actors to spoof the identity of users.

“While they might be able to obtain passwords from the dark web, hackers will have a very hard time imitating every behavior and habit that the targeted user has.”

Read more here.