India must understand the need for data anonymisation and also build the idea of consent when dealing with facial identification technology
The pandemic pushed many enterprises into survival mode. They started exploring new technologies like computer vision to help them work efficiently and safely. Computer vision uses live feed from CCTV networks to generate actionable insights in real time. Many offices have started contactless attendance wherein a camera identifies a person even with a mask, using facial recognition technology. There is also an increase in the use of facial recognition for contact tracing to stop the spread of Coronavirus from an infected person. However, the spread of this technology globally also brought to light its misuse, such as racial discrimination. After several cases of bias against people of colour, the US Democratic lawmakers introduced a Bill that would ban the use of facial recognition technology by federal law enforcement agencies.
India hasn’t implemented a countrywide facial recognition programme yet. We are in the perfect stage wherein we can learn from the mistakes of others when it comes to creating laws that will monitor this powerful technology so as to protect our citizens’ right to liberty and privacy. Like any emerging technology, there are multiple applications of facial recognition that appear completely harmless. For instance, Facebook tags users in pictures and Google Photos organises them according to faces. Interpol has been using the technology to make the immigration process safer, easier and faster. It uses information from 160 countries to prepare a criminal database. Over 650 missing people or “persons of interest” were reportedly recognised using this technology in 2016. Today, it is being used across multiple sectors. Privacy and surveillance concerns arise due to multiple cases of racial discrimination and illegal use of information. An innocent man in Michigan was forced to spend one whole day in police custody after a facial recognition system mistakenly identified him as a shoplifter. There have been numerous instances of facial recognition systems wrongly identifying people of colour. Various algorithms provide incorrect matches for darker skinned individuals.
So what does the US Bill say? In the light of several protests and the Black Lives Matter movement, the Bill prohibits the use of facial recognition technology and its federal funding. It will be illegal for any federal agency to use or have access to surveillance technology within the geographical boundary of the US. Furthermore, the information obtained in any form, resulting in the violation of this Bill, cannot be used for judicial proceedings. It clearly states that this technology cannot be used without a federal law and relevant provisions to ensure maximum security. Incidents of racial discrimination have always been a major concern in the US and the killing of George Floyd in May put the spotlight on errors in facial recognition technology, too. Various studies have shown that it does not have the same level of accuracy for people of different skin colours and gender. It works best when identifying white men but falters the maximum when faced with images of dark-skinned women. Wrong identification by algorithms has led to increased cases of police brutality and racism.
Does India need a counterpart? There are people of varying skin tones in our country but in the sphere of criminal justice, the discrimination is not based on their colour. We discriminate on the basis of caste, tribe, economic status and religion. Hence, in India, the facial recognition software is not a threat in the same way it is in the US. However, fears over privacy that arise out of the misuse of this technology are valid for us because as of now, there is no law protecting the privacy of people. On a trial basis, the Central Board of Secondary Education and the Indian Railways became two institutions to successfully use facial recognition on a large scale. The railway station at Bengaluru, too, is set to get an AI-powered CCTV system that uses facial recognition technology to aid in creating and maintaining a database of criminals. The problem with such a powerful technology is that it needs supervision at every stage. People in India have not put much thought into what facial recognition could do for us or the threats it could give rise to. Some believe it is the future and some are aware of data protection and misuse of information.
The way forward: Though the advantages of technological advancement are undeniable, it is vital to take into account situations of unprecedented social insecurity that this brings. We need to address the concerns about data privacy, protection and ethics in order to create a strong legal framework to guide the workings and usage of biometric and facial recognition software. As a first step, the Personal Data Protection Bill 2019 needs to be reviewed and made into an Act. A facial recognition algorithm picks up the biases of the data fed by people into it. People working with facial recognition algorithms should be careful when feeding data and also when deriving insights out of it. We must understand the need for data anonymisation and also build the idea of consent when dealing with facial identification technology. Constant improvement should ensure zero scope of error eventually. With the right policies and laws in place, facial recognition will ensure our kanoon (law) is not andha (blind) anymore as enforcement agencies will be able to see 24 frames per second through the millions of CCTVs across the country. Hopefully for our security.
(The writer is Co-Founder and CEO of Integration Wizards Solutions)
