Automated Face Recognition Technology (AFRT) captures people's facial traits and shapes to prepare a database for prospective comparison. AFRT is a development of facial 'profiling' or 'mapping,' utilised in criminal justice systems worldwide since the nineteenth century and is still employed today. Automated Facial Recognition is a biometric analysis that identifies or verifies patterns derived from an individual's facial characteristics and features. India is one of the countries where government surveillance has become a pressing problem in terms of data privacy. Regulations that allow governments to access citizens' data undermine some countries' comprehensive privacy safeguards of their residents.
Along with countries with a high level of government monitoring, such as China, India has been recognised as having few constraints regarding data privacy and protection. Government surveillance is a subject of caution. The direct application of such technology is not legally authorised. As a result, a detailed legal framework enacted by the Indian Parliament that enables the development and maintenance of such technology is required. Currently, there is no explicit regulation in India that allows for the deployment of these technologies. The Information Technology Act, 2000, India's primary legislation on electronic format, is devoid of any mention of facial recognition.
AFRT can be used for one-to-one matching, verification of an individual's identity, or one-to-many database searches. Facial recognition as a technology has its own set of advantages. It enables law enforcement organisations to track not only offenders but also missing children by comparing facial traits. In India, it has been used in pilot projects. This technology can also be utilised at airports to demonstrate the most efficient method to be employed for the purposes that can be met in airport construction. However, the accuracy of AFRT has been a critical point of disagreement.
One of the pivotal moments in this argument occurred when an academic group from MIT and Stanford University conducted a ground-breaking experiment to examine Amazon's facial recognition technology, Rekognition.
The scenario was exacerbated by the fact that when women with darker skin tones were subjected to facial profiling, the mistake rate was much higher. Furthermore, because law enforcement organisations were using this technology in several states across the United States, the discourse took on a solid undertone of racial injustice and African-American vulnerability, mainly due to the employment of such technologies. Nevertheless, despite its flaws, AFRT is widely used by law enforcement agencies around the world. For example, numerous state police forces in India have already begun using various forms of AFRT in their surveillance and monitoring functions.
Ethical and Legal Concerns related to AFRT
While AFRT is increasing worldwide, India's lack of data protection and privacy law leaves the country unprepared to exploit the technology. "There is no legislation in India to ban the misuse of facial recognition technology," Pavan Duggal, one of the country's leading cyberlaw experts, said in an interview. The IT Act does not explicitly address this technology's misuse. There is also no blanket ban on using this technology, maybe due to the benefits that could arise from the correct application of the technology, which drastically reduces the time required for recognising persons or objects in images and videos.
During four-day testing of a facial recognition system, Delhi Police could locate nearly 3,000 missing children. However, in India, anyone can abuse this technology without fear of repercussions from the law. "Here, the law has failed to safeguard the citizen," he says, adding that self-regulation of facial recognition will be ineffective. "The sooner we can develop an efficient legal structure to control facial recognition technology, the better it will be for the country and its inhabitants," Duggal further said.
Furthermore, several privacy issues are being raised by the citizens of India, which at the same time were well supported and backed by the wordings of cyber security experts. For example, the UIDAI declared in January 2018 that it would enable AFRT as one of the ways of authentication under the Aadhaar Act. The UIDAI had also compelled telecom service providers to begin face authentication of their subscribers in future circulars. While it is no longer possible for the government to mandate Aadhaar-based face authentication by private entities such as banks and telecom companies following the Supreme Court's verdict in the Puttaswamy case, the possibility of it being used by the government for the distribution of welfare benefits remains legitimate.
FRT through the Puttaswamy Lens:
The court established that while privacy is not an absolute right, any interference in an individual's private life should only be done in a "fair, just, and reasonable" manner. The Supreme Court of India, in the landmark decision in the Puttaswamy case in August 2017, declared that privacy is a fundamental right under the Indian Constitution. While this right is not expressly recognised in the Constitution, the court determined that it derives its fundamental status from its interaction with several other constitutional rights. However, on these bases, the legitimacy of the technology suggested that the NCRB tender has already been called into question in a legal letter sent to the NCRB and the Ministry of Home Affairs. The Internet Freedom Foundation emphasised the lack of any statutory foundation for establishing such a system. Therefore, it breaches the first test established by the Supreme Court in the Puttaswamy case.
Furthermore, the notice criticises the planned system for allowing photographs of persons to be collected without their knowledge or consent, being vulnerable to misidentification and discriminatory profiling, and lacking proportionality measures and an oversight mechanism. The government argued in the Aadhaar dispute that a person's face is a widely accepted form of identity. Therefore, we commonly submit facial images for driving licences, passports, school admissions, et cetera.
On this basis, the state asserted that there is no "reasonable expectation of privacy" in facial pictures. The court rejected this argument, saying that when state agencies acquire and maintain demographic or biometric data, it warrants a more profound look utilising the proportionality standards.
In the context of AFRTs, it is worth noting that the processing of a facial image to develop an individual's unique facial pattern, as well as the ability for that data to be stored and processed later, lends facial recognition a distinct personality that differs significantly from putting one's face or photos in the public domain. As a result, the argument that people's faces are available in the public domain should not be considered a valid mitigating element in deploying AFRTs for law enforcement reasons.
Cyber security is a crucial component and element of facial recognition technologies. When automated facial recognition systems are applied, large data sets containing personally identifiable information will be generated. In order to defend people's rights and liberties, it is necessary to ensure the cyber security of such data. Law enforcement agencies all across the world are wrestling with the limitations and hazards of facial recognition.
Pros of AFRT:
There are numerous advantages that facial recognition may provide society, ranging from minimising unnecessary human interaction and work to preventing crimes and boosting safety and security. In rare cases, it can even aid in medical attempts.
It aids in the search for missing people and the identification of criminals.
By comparing faces on live video feeds with those on a watch list, law enforcement organisations can identify criminals who have no other means of identification and find missing people. They have even used it to track down lost children, combining face recognition with ageing software to forecast how children will look in the future and find them even if they have been missing for years.
It safeguards businesses against theft.
Facial recognition software can be an effective deterrent against shoplifting. Business owners use the software and security cameras to identify known or suspected thieves, and the presence of the cameras themselves works to dissuade theft in the first place.
It cuts down on the number of touchpoints.
Unlike other types of security procedures, such as fingerprints, facial recognition requires fewer human resources. It also does not necessitate direct human connection or physical contact. Instead, it employs artificial intelligence to make the process automatic and frictionless. It also restricts touchpoints when unlocking doors and smartphones, withdrawing cash from an ATM, or doing any other action that usually necessitates a PIN, password, or key.
Cons of AFRTs:
As with any technology, there are possible negatives to employing facial recognition, such as threats to privacy, abuses of rights and personal liberties, data theft, and other crimes. There is also the potential of errors owing to technological defects
It allows for the commission of fraud and other crimes.
Criminals can also utilise it to commit crimes against innocent victims. For example, they can acquire personal information, including imagery and video collected from facial scans and kept in databases to perpetrate identity fraud. A thief may use this information to obtain credit cards and other debt, open bank accounts in the victim's name or even build a criminal record in the victim's name.
It imposes on personal freedom.
Being videotaped and scanned might give people the impression that they are constantly being monitored and scrutinised for their actions. Furthermore, police can use face recognition to run everyone in their database through a virtual criminal lineup, equivalent to classifying one as a criminal suspect despite the lack of probable cause.
Technology can be deceived.
Camera angles, lighting levels, and image/video quality are all elements that can impair the technology's capacity to identify faces. In addition, people who wear disguises or make minor changes to their look can also fool the software.
Facial biometric data is one of the most sensitive types of personal data. Therefore, any deployment of this technology by state agencies or the private sector must be preceded by implementing a solid data protection law. Such usage would still violate the Supreme Court's standards in the Puttaswamy case regarding the right to privacy. Moreover, the proposed data protection framework would not provide the level of responsibility required from the diverse stakeholders involved in implementing AFRTs.
The use of facial recognition and other biometric identification systems can be justified for specific security goals. Thus, privacy and other issues can be minimised. It is, nevertheless, critical that its use be subject to accountability measures to ensure against misuse. Therefore, citizens should be informed about biometric facial recognition systems, and consent must be obtained for their usage for specific and justified purposes. Furthermore, their usage should be debated publicly, supported by legislation, and subject to procedural fairness.
foxeslawjournal.com 2020. Legal Challenges in Implementation of Automated Facial Recognition Technology in India. Accessed 10 September 2021 Available at: https://www.foxeslawjournal.com/post/legal-challenges-in-implementation-of-automated-facial-recognition-technology-in-india-viola-rodrig
Buolamwini and Gebru (n 9).
The research showed that Amazon's Rekognition tool misidentifies darker-skinned women as men 31% of the time; lighter-skinned women were incorrectly identified 7% of the time, and darker-skinned men had an error rate of 1%. See Hardesty (n 9).
Mehrotra, K., 2021. Indian faces were run through facial recognition tech tools. Here is why one should be concerned. Scroll.in Accessed 10 September 2021 Available at: https://scroll.in/magazine/1001836/facial-recognition-technology-isnt-wholly-accurate-at-reading-indian-faces-find-researchers
Mint. 2019. Is India prepared to tackle misuse of face recognition technology?. Accessed 11 September 2021 Available at: https://www.livemint.com/technology/tech-news/is-india-prepared-to-tackle-misuse-of-face-recognition-technology-1549646571603.html
Justice KS Puttaswamy (Retd.) and Anr v. Union of India and Ors (2017).
datagovernance.org 2019. Adoption and recognition of Facial Recognition Technologies in India. Accessed 11 September 2021 Available at: https://www.datagovernance.org/files/research/NIPFP_Smriti_FRT_-_Paper_5.pdf
Gargaro, 2021. The pros and cons of facial recognition technology. [online] Itpro.com. Accessed 9 September 2021 Available at: https://www.itpro.com/security/privacy/356882/the-pros-and-cons-of-facial-recognition-technology