FACIAL RECOGNITION: THE RISING MANIPULATION
Author: Ankit Kumar, IV year of B.A.,LL.B. from Army Institute of Law
In today's fast-paced world, where technology is blooming at its full potential, the new technology can prove itself to be a double-edged sword. We have all sorts of laws and regulations for civil, criminal, and many other matters but there are new aspects that are yet to be explored, one of them being the legal framework to govern Data Protection and IT issues.
This is one of the serious concerns for the regulators of law as modern technology with all its benefits can also be put into use for various malicious activities most of which are beyond apprehension to the common people. At current, we cannot deal with technology as a whole but the least we can do is focus on the problems at hand.
One such pertinent problem is the rapid deployment and use of the Automated Facial Recognition System (hereafter 'AFRS') in India.
AUTOMATED FACIAL RECOGNITION TECHNOLOGY
In earlier times, a simple version of facial recognition was used which was based on key facial features and geometry for designing a facial map which then referenced with millions of images to identify similar faces. But AFRS is different, it is based on a type of artificial intelligence in which neural networks are trained on millions of pictures and other data collected from different sources. Hence, it's more advanced and is capable of recognizing facial data from real-time CCTV footage.
These capabilities enable law enforcement in identifying potential perpetrators and prevent crime and thanks to their speed and efficiency it's being used all across the world by various governments and agencies.
But this technology is also not bereft from various problems; there are various concerns about its accuracy and privacy. Also, there are apprehensions of an Orwellian state where big brother is watching your every move meaning abuse of the technology by governments and hence dangerous to democracy as a whole.
ACCURACY OF AFRS
One of the primary concerns about AFRS is its accuracy. In order to clear this out, Stanford University in collaboration with MIT conducted research on Amazon's facial recognition tool 'Rekognition'.
The research found that in spite of all the sophistication in designing this tool, it is not devoid of problems. They found that inaccuracy and error rates are significantly higher when this tool is subjected to images of minority and dark-skinned women for facial profiling.
These researches brought hostility among the African-Americans, as in the US many law agencies are using this technology. Also, a similar issue arose in England, and there has been lots of protest against the use of AFRS by the London Police.
The dilemma is that with all its inaccuracies countries all across the world are imposing bans and regulations here in India we are introducing and deploying this technology. Recently EU also proposed for ban on FRS in public spaces as AI is found to be biased towards people and is dividing them based on ethnicity, gender, political or sexual orientation.
AFRS IN INDIA AND ITS REGULATIONS
The National Crime Records Bureau ('NCRB') has recently proposed establishing National Automated Face Recognition System while many states in India are already at different stages of its deployment. At present there are 16 different FRS are in active use by various central and state governments for security and surveillance. In addition to this, we still lack a proper data protection law which may help in clarifying the position of AFRS, hence making things more complicated. It's very unfortunate that where the technology is being introduced all across the law and regulations will still take a while in catching up and hence is left in the hands of enforcement agencies.
The issue of AFRS is similar to the narco analysis problem as it was also unregulated and left to law enforcement to decide. It was challenged multiple times as it may potentially be dangerous to innocents and hence in Selvi v. State of Karnataka[i], the apex court held that narco tests, as well as other involuntary mechanisms like brain mapping and polygraphs are unconstitutional, unless the accused consented to them.
It was also found that Delhi police was using AFRS for identifying and comparing facial data of people involved in violence during the protests against the anti-citizenship act in Jamia Millia Islamia with criminal datasets.
REGULATORY CONCERNS IN INDIA
Some further concerns about AFRS are discussed below:
1. Lack of transparency
There is a lot of ambiguity in the NCRB's decision of establishing a nationwide facial recognition system.
The basis on which the AFRS system is established in India is said to be the cabinet note of 2009 but a cabinet note is not a legal substance and hence the establishment is legally invalid.[i]In addition to this no details related to its use is made public like which datasets are going to be used for harnessing etc.
This lack of information causes distrust among citizens, hence it's necessary to make the relevant information public. Also the Right to Information, 2005 warrants such information to be accessible to citizens.
2. Threat to Right of privacy
According to this test, the interventions made by the state must be established in accordance with the law, in pursuit of state interest, and should prescribe checks and balances to prevent abuse of the State's Surveillance powers. From this, it can be figured out that AFRT lacks two main pillars that are in accordance with law and checks and balances.
As India is a country of diversity containing various different facial structures and with a huge population, so even a slight margin of error can result in chaos and tyranny among the population. Also, such technology without any defined legal framework and data protection regulations makes citizens more vulnerable to privacy abuses. Hence there is a strong need for a legal framework to maintain law and order.