By Tripti Jaiswal, First-year Student of Hidayatullah National Law University, Raipur, Chhattisgarh.
Introduction
An average human can recognize faces very well. So can your shiny new smartphone. The ability to perceive countless faces is important in our everyday life, and the automation of this cognizance has facilitated us with the current facial recognition technology. It can recognize faces quicker than the blink of an eye, and it doesn’t falter at people’s names like us common folks. Now that computer scientists have engendered tools that can turn faces into name tags, it’s worth reflecting on how we got here and what we stand to lose.
The concept of Facial Recognition is not new; the earliest example of facial recognition can be traced back to the 1960s. Technology for detecting and recognizing facial images has gotten better with time and has gained accuracy. Meanwhile, the invention and popularization of digital photography and social internet paved the way to create massive databases for the algorithms to learn from. Millions of pictures were not solely posted, but also conjointly tagged by the users on Facebook alone.
This data is perpetually being employed to idealize the technology. The utilization has endless possibilities: ranging from the creation of enticing features like Face ID in Snapchat to selective product advertisement and scrutinizing surveillance. These algorithms learn to detect and recognize facial images by matching millions upon millions of photographs to their similar-looking counterparts. Yet, due to the presence of gender and racial diversity in the samples used; they might not perform successfully on all types of individuals. Facial recognition or facial apperception tech has been shown to be inaccurate in distinguishing darker-skinned women, women from ethnic minorities, and transgender people. Amazon came under fire when U.S. congressional lawmakers expressed concerns about its facial recognition tool, ‘Rekognition’ due to its inherent bias towards dark-skinned people[i]. That hasn’t obviated it from being packaged and distributed as ready-to-use software.
While the remote inaccuracies seem trivial when applied to social media, it is crucial to remember that these databases will be used by law enforcement authorities soon. India has already been preparing to launch a nationwide facial recognition system, AFRS, in an effort to improve police investigations, with the aid of a nation-wide centralized database of images.[ii] During a trial last year, the police claimed to have identified approximately 3,000 missing children using the technology.
Many states like Delhi and Telangana have started using facial recognition to monitor the public during protests, demonstrations, and public gatherings.[iii]
Rights of children and the elderly
Children, who are most likely to grow and mature, have varied facial factors as compared to other age groups. Aging negatively affects the precision of Facial Recognition technologies, especially in children and older people. Due to the particular susceptibility of children, there are considerably more false negatives of them as compared to other age groups. Due to their rapid growth, their facial images need to be used for preserving their identity. Facial recognitions are used by the police to help trace missing and abducted children.
Therefore, careful consideration should be taken while processing images of children and older persons, so that they are not disproportionally affected by the negative consequences of this technology.
Implications of using live facial recognition
The facial images of biometric data can be categorized as sensitive information and can be easily captured in public places. Since people whose pictures are taken and processed might have no idea that this is happening, they have little chances of challenging or disputing in case of possible misuses.
The deployment of Facial Recognition Technology should be done while respecting the private lives of the people and safeguarding their personal data so obtained. It’s one of the most classic values that need to be instilled in the process of installation of this tool. While India does not have a dedicated law on data protection and privacy[iv], Section 21 of the Indian Constitution protects the right to privacy as a fundamental right. Live facial recognition involves the biometric processing of faces captured in a public place for possible identification of individuals on a ‘watchlist’, and the data thus collected is retained to the database. The collection of facial images for the purpose of facial recognition is not yet governed by any stringent state laws. Since facial recognition, both with data pulled from publicly available resources as well as live surveillance footage has been coherently useful only recently, the laws for the protection of individuals have been lagging haplessly.
Another subject that needs to be broached is whether this sensitive information can be kept safe by the government. While AFRS is strictly accessible to law enforcement agencies, data leakages could send this information in the hands of people with ill intent.
Even immensely colossal tech companies, which promise total data protection, can be targeted by hackers imperiling personal information. Facebook, infamously, has had multiple data breaches in the past.[v] Avoiding potential data leakages to ensure the protection of biometric data and thus, the privacy of common individuals is going to be arduous. While facial recognition software is on the rise, increasing data security should be the foremost concern. This requires measures to be put in place to integrate safeguards for preserving sensitive information.
Right to Equality has been granted to all citizens by the Constitution and is embodied in Articles 14-18. The principles of equality and non-discrimination have been salient features since the very inception of the constitution. The deeply complex decision-making algorithms of facial recognition are still flawed. It is capable of inaccuracies that are disproportionately higher in darker-skinned people.
While algorithms will improve with further development of technology, the biases incorporated in the system are present in the algorithm as well as when officers decide what action to take following a match.
Criticisms have been made on the increased surveillance of protests and rallies by law enforcement authorities. While the general purpose is reasoned to be identifying violent elements in case of riots, it negatively affects the freedom of expression and association. If people are being watched by facial recognition tech in public spaces, especially during protests, it creates a chilling effect; often discouraging them from attending demonstrations, thus, taking away their freedom to a certain extent as the surveillance could make them a target. This seriously interferes with the provisions of Article 19 of the Indian Constitution. These kinds of technologies should be used towards satisfying the needs of national security, and not to enforce a system of social control, like in China.[vi] It is therefore vital to establish that facial recognition is not used to target potential targets; the technology needs to be used transparently and with the aid of robust legislation. In this regard, a higher administrative court inMünster, Germany has declared that the police can no longer publish the pictures taken at demonstrations or public gatherings through social media or any other public platforms including police media channels due to its negative effect on the freedom of association and infringement of the right to assembly.[vii]
Using facial recognition technologies – a technology that is still developing- affects a range of fundamental rights for a multitude of reasons. The way facial images are obtained and processed – potentially without consent or opportunities to opt-out affects people’s dignity and the intrusive ways of procuring facial images is an invasion of privacy.
[i] Letter Congress of the United States to Jeffery Bezos (CEO, Amazon) (Nov. 29, 2018) https://www.markey.senate.gov/imo/media/doc/Bicameral%20Amazon%20Recognition.pdf(last visited on Jul 16, 2020). [ii]India setting up world’s largest facial recognition system (2019), https://www.dw.com/en/india-setting-up-worlds-biggest-facial-recognition-system/a-51147243 (last visited on Jul 16, 2020). [iii] Anthony Kimery, India set to stand up world’s largest government facial recognition database for police use (March 11, 2020), https://www.biometricupdate.com/202003/india-set-to-stand-up-worlds-largest-government-facial-recognition-database-for-police-use (last visited on Jul 15, 2020). [iv]Kochhar& Co., Data protection and privacy in India (August 27, 2019), https://www.lexology.com/library /detail.aspx?g=d1edde8f-71b9-49cb-b333-35fcae73402b (last visited on Jul 17, 2020). [v]ShwetaGanjoo, Facebook faces another data breach, data of 267 million users exposed, INDIA TODAY TECH, Dec 20, 2019, https://www.indiatoday.in/technology/news/story/facebook-data-of-267-million-users-exposed-online-1630084-2019-12-20 (last visited on Jul 18, 2020). [vi] Lauren Dudley, China’s Ubiquitous Facial Recognition Tech Sparks Privacy Backlash, THE DIPLOMAT (March 7, 2020), https://thediplomat.com/2020/03/chinas-ubiquitous-facial-recognition-tech-sparks-privacy-backlash/ (last visited on Jul 18, 2020). [vii]Police in North Rhine-Westphalia can no longer publish photos of protests, a court rules (2018) https://www.dw.com/en/police-in-north-rhine-westphalia-can-no-longer-publish-photos-of-protests-a-court-rules/a-50463890 (last visited on Jul 18, 2020).
Comments