GUEST ESSAY: The privacy implications of facial recognition systems rising to the fore

By Lance Cottrell

Tech advances are accelerating the use of facial recognition as a reliable and ubiquitous mass surveillance tool, privacy advocates warn.

A  string of advances in biometric authentication systems has brought facial recognition systems, in particular, to the brink of wide commercial use.

Related: Drivers behind facial recognition boom

Adoption of facial recognition technology is fast gaining momentum, with law enforcement and security use cases leading the way. Assuming privacy concerns get addressed, much wider consumer uses are envisioned in areas such as marketing, retailing and health services.

According to Allied Market Research, the facial recognition systems market is in the midst of rising at a compounded annual growth rate of 21% between 2016 to 2022. The research firm projects that the facial recognition market will climb to $9.6 billion by 2022.

Pieces in place

Ntrepid is focused on the privacy ramifications associated with these developments. As privacy concerns get addressed, facial recognition technologies are expect to emerge as a consumer favorite, when compared to other biometric authentication systems, such as voice, skin texture, iris and fingerprint systems.

This trend is rapidly unfolding because all of the required pieces are finally in place. Cameras have become cheap and ubiquitous. We all take far more pictures (mostly of ourselves) than ever before in history. We are also putting cameras everywhere, in devices, in our homes, in our shops, on our streets.


At the same time, we are using social media and other tools to put this enormous trove of pictures of faces up in the cloud where they can be analyzed. Many of these photos are tagged with the identity of the people in them.

We are also seeing tremendous progress in the science and machine learning required for the recognition of faces in those photos, in part powered by the huge amount of training material available.

Security use cases

The combination of these trends makes the widespread adoption of the technology at this time almost inevitable.

Facial recognition will be used increasingly both for identifying people out in the world as well as for powerful biometric authentication for our own security.

Like any technology, facial recognition is a double edged sword. Facial biometric security is both strong and convenient. It can protect access to devices or buildings effectively and conveniently.

Law enforcement often cites the ability to track terrorists through a city as an argument for wide spread surveillance with facial recognition. This wider application of the technology also means that we are losing any expectation of privacy while in public.

It has always been true that we are tracked at all times when online, but for practical purposes we have been anonymous when walking down the street. Going forward, we can expect that someone, maybe many different entities, will be able to identify and track us at all times.

Avoiding Orwell

This kind of Orwellian scenario is already developing in China. Because the technical winds are strongly behind this trend, the only way to limit it will be through legislation and regulation. One solution for online facial recognition would be to trick the systems to allow the use of subtly altered real photos so that the subject will not be recognized by automated systems.

This kind of technology would have many applications. On a consumer level, it might allow you to post somewhat inappropriate party photos while preventing them being tagged with your identity.

The development of universal video surveillance with facial recognition is generally happening away from the public eye. There is very limited visibility into the capabilities and limitations of the solutions, how they are used, and whether they are consistent with the kind of society we want. It is critical that these conversations and tradeoffs take place now before full deployment and use of the technologies is a fait accompli.

About the essayist: Lance Cottrell is chief scientist at Ntrepid, which supplies endpoint security and information management solutions.


Share on FacebookShare on Google+Tweet about this on TwitterShare on LinkedInEmail this to someone