FAIFE is marking the 20th anniversary of the IFLA Statement on Intellectual Freedom. As part of this, Jonathan Hernandez-Perez, a FAIFE member from the National Autonomous University of Mexico (UNAM) has shared the below blog on the subject of facial recognition technology, and what it means for libraries and their values.
Facial recognition is one of the current obsessions of the tech industry, with regular high-profile product launches meaning that it is also high on the public agenda. It has developed rapidly over the last years, making it possible to undertake tasks that usually take hours in just the blink of an eye.
Yet as the number of public spaces that use this technology keeps rising, so too do the public’s concerns about privacy and surveillance, leading to many more negative media headlines and an intense social media debate. This blog explores what facial recognition technology is, the questions it raises, and what this means for libraries.
What is Facial Recognition?
Facial recognition is a type of technology that allows the verification and identification of a person through analysis of his/her facial features. This technology has been with us since long before the coming of the internet.
With the intention of obtaining a definition of the “criminal face” during the 19th century, several facial patterns of ex-convicts and criminals were gathered. Fortunately, the idea that the measurements of someone’s head are associated with criminal tendencies has long been rejected.
However, some of the techniques involved have been enriched and improved, involving a greater number of actors and interests, leading to current technology that makes our daily lives more comfortable, from the basics of unlocking our cellphones or automatically “tag” a friend in a picture, to the more complicated issues, such as airport check-ins, tools to validate our identities at ATMs, or even means of gauging emotional responses. It turns our face, our emotions, and expressions into a bar code.
Furthermore, facial recognition has the potential to be combined with other technologies in order to combine and enhance the tracking that happens in the digital and physical sphere.
Enabling Surveillance, Hidden Bias
The convenience allowed for by facial recognition comes with a price, and in the digital era the cost is our privacy. This is because nowadays, our facial expressions – the very essence of human social and emotional interaction – have become an object of experimentation, propaganda, and database development. Arguably, we are only partially aware of the extent and consequences this technology could have in a very short time period, particularly because biometric technologies are still not widely understood.
A particular worry is the degree to which facial recognition technology enables mass surveillance. In 2013 the IFLA Trend Report stated that expanding data sets – for example of faces – held by governments and companies will support the advanced profiling of individuals, while sophisticated methods of monitoring and filtering communications data will make tracking those individuals cheaper and easier, warning that serious consequences for individual privacy and trust in the online world could be experienced. This now appears to be coming true.
In 2014, Insecam demonstrated the possibility of illicitly obtaining images from security and surveillance cameras that use weak passwords. This poses a particular threat to public privacy since they are placed in public spaces. Meanwhile, in 2016, a Russian photographer carried out an experiment to show how easy it was to identify strangers in the streets using only one picture to identify them. More recently, FaceApp, which takes your photo and gives an idea of how you’ll look decades from now, put back into focus the privacy vulnerabilities of mobile applications.
The consequences of the implementation of facial recognition technologies have come into the spotlight with the recent protests in Hong Kong, showing how our faces can become a weapon either for persecution or prosecution. Responding to public pressure, some cities have begun to ban the use of facial recognition software by state agencies; San Francisco, Somerville, and Oakland are the first cities in the United States with a regulatory law over this topic.
A further concern is around the risk of bias in facial recognition technologies. These systems are usually trained on a different number of faces from specific groups of people with similar facial characteristics (Mostly Caucasian) which could lead to the failure of people recognition in a more diverse environment, and in a legal way, this could lead possible mistaken identification entailing people to crimes they didn’t commit.
This matter involves race, gender identity, and sexual orientation issues which makes it more threatening and harmful, there are a number of examples of how this technology is developing an automated racism.
Impacts for Libraries
This year IFLA celebrates the 20th anniversary of the “IFLA Statement on Libraries and Intellectual Freedom”. It is as crucial as ever to underline one of its key principles:
Library users shall have the right to personal privacy and anonymity. Librarians and other library staff shall not disclose the identity of users or the materials they use to a third party.
This principle is relevant at the moment because today, privacy and mass surveillance are some of the most pervasive and threatening issues we face. Certainly, we risk seeing facial recognition turn from being a “fad” into a normal practice and would eventually be part of a new common sense and part of our mainstream culture. This would imply an important loss of privacy.
Libraries have always worked to keep up to speed with new technology and to make best use of the possibilities it offers. Therefore, facial recognition will also impact in their services.
Facilitating the registration, loan, and access to information resources could be a very attractive reason to implement this kind of technology in libraries. Companies are already selling biometric software for book loans and some libraries have been using these systems for a couple of years now.
In the near future, libraries may be able to offer material based on our facial expression, then, our face could become a personal card that does not belong to us, associated with all the data about books read, web pages consulted, and topics we are passionate about. If we are not aware of the extent of this information, it could become a big threat to our privacy.
As a result, the use of this technology in libraries is a matter that should be analyzed in the light of user freedoms and rights, and the potential damage it could do to privacy and intellectual freedom, values that libraries have defended for years.
Therefore, libraries must provide digital secure spaces where our movements are not tracked and develop privacy programs for librarians and their community. An interesting example is the Library Freedom Institute, which teaches librarians and patrons how to protect their privacy online and how to influence public policies on this matter.
Although we may share similarities with other people all over the world, every face has its own interesting and unique features. Thousands of databases are daily fed with biometric information and we are taking part into this dynamic through our daily digital behavior. But the problem shouldn´t be attached to the user. Knowing the value of our data or agreeing on the terms and conditions companies impose it’s not enough, neither is derision or banning some apps or software.
What is required is having strong legal frameworks and policies that protect individual rights for limiting such tracking. Libraries can both lead the way in their practice, and push for the right laws and regulations in their advocacy.