In Facial Recognition We Trust?

Tulsi Parida
4 min readOct 9, 2019

--

New technology is often initially met with scepticism, with a small number of innovators and early adopters paving the way for the majority to eventually move beyond their initial reservations. What causes these innovators, early adopters, and then eventually the public to adopt new technologies?Having confidence in something new requires trust. When people overcome key trust barriers, they can leap into the unknown and feel more comfortable using new technologies that once seemed alien.

However, sometimes, new innovations are discarded because the trust barriers are insurmountable. Innovation in and adoption of automated facial recognition (AFR) surveillance cameras can be linked to three key trust barriers: fear of big brother, tension between privacy and security, and a general fear that technology will make costly mistakes. I argue here that perhaps these trust barriers are too high for AFR surveillance cameras to be worthy of our trust at all.

Big Brother

Jeremy Bentham, an English philosopher of the 18th century, introduced the panopticon as a type of prison system designed so that prisoners could be watched without knowing when or by whom. This idea of central inspection was later elaborated upon by Foucault as a cautionary tale about asymmetrical government surveillance. Foucault posits about the prisoner that “he is seen, but he does not see; he is an object of information, never a subject incommunication”. The prisoner here has no power and the guard or government has all information and therefore all power.

Media outlets often use the term “big brother” when discussing law enforcement’s usage of AFR inciting (perhaps a rational) fear of the technology. Concerned individuals might try to opt out of having AFR used on them. There has already been a case of a British man trying to cover his face as he passed by an AFR surveillance camera.

Who exactly is the customer of AFR surveillance cameras and while this could be in the best interest of law enforcement agencies, is the practice in the best interest of the general public?

Privacy vs. Security

While AFR surveillance cameras have been touted as a new security measure, many civil liberty groups have been citing it is an infringement on individual privacy, the next key trust barrier. People in most western democracies believe they have a right to privacy and seeing AFR in direct opposition to that right translates into a sometimes insurmountable trust leap. A similar concern arose when regular CCTVs were installed all around the UK. The country saw an explosion of security cameras in the mid-to-late noughts with a simultaneous uproar about privacy surrounding it.

Today, however, most people have become accustomed to CCTV. Most citizens have accepted CCTV cameras as an inevitable part of their public life and as a tool to help keep public areas safe. I am doubtful that this same shift will occur with AFR surveillance cameras because of the current regulatory environment. Regulation around AFR surveillance camera use specifically for law enforcement has recently become a topic of contention in major cities (and countries) around the world. San Francisco, most recently, implemented a ban on the use of AFR technology for law enforcement.

Could this technology be misused to control the population in a way that goes against core democratic principles of privacy and agency?

Costly Mistakes

The final key trust barrier for AFR surveillance cameras is that errors in the technology could be considered too risky because of the potential of grave consequences. This is not a challenge unique to AFR surveillance — when the telephone was a new invention, for example, people were afraid of the potential costly consequences. People were worried that if the technology went wrong, they would get electric shocks. They were also worried that people might spend too much time gossiping on phones, changing social behaviour.

Similarly, people are afraid that if AFR technology goes wrong, police will misidentify criminals (as has already happened) thus potentially causing a rift between law enforcement and the general public. The trust barrier here is mistakes that are too costly. There is also a fear that if the data falls into the wrong hands, it can be weaponized.

Are the most vulnerable groups going to be harmed more by this than others?

AFR surveillance cameras have been piloted all over the world, but the general sentiment publicized by the media is that AFR surveillance cameras as they operate today cannot be trusted. This has been true of other innovations in history, but I argue here that this lack of trust in AFR surveillance cameras specifically actually has little to do with the technology itself, but has everything to do with the institution deploying this technology. Because of a deeply entrenched and widespread lack of trust in the institution of law enforcement and the subsequent reinforcement in that lack of trust by regulations such as the ban on AFR for law enforcement in cities like San Francisco, AFR surveillance may not be deemed a trustworthy new technology.

However, the lack of institutional trust that could prevent innovation and adoption of AFR surveillance cameras does not translate into other types of AFR technology. The field of healthcare, for example, has received a tremendous amount of positive press for its use of AFR . Since there is not the same scepticism of healthcare professionals as there is of law enforcement, AFR for healthcare could have a lot more buy-in from the general public, and a subsequent explosion of innovation in the space.

In the case of AFR, the institution using the technology actually matters more than the technology itself when it comes to trust, and when the institution behind it is deemed trustworthy, public scepticism will diminish.

--

--

Tulsi Parida

I head data solutions for public sector at Visa. I care about intersectional feminism and inclusive tech.