Facial recognition technology (FRT) is a rapidly emerging field, with applications spanning from security to law enforcement. However, this technology also raises several questions regarding privacy, accuracy, and its potential misuse. A recent revelation by the Internet Freedom Foundation, a digital rights organization based in New Delhi, has ignited a new debate about FRT’s use by the Delhi Police.
Delhi Police’s Use of Facial Recognition Technology
According to Right to Information (RTI) responses, the police department treats matches above an 80% similarity generated by its facial recognition system as positive results. Matches that fall below this threshold are categorized as false positives requiring additional corroborative evidence. This leaves concerns about potential misidentifications and the targeting of over-policed communities.
The Delhi Police gathers the matching photos and videos under Section 3 and 4 of the now replaced Identification of Prisoners Act, 1920, currently the Criminal Procedure (Identification) Act, 2022. This act has sparked fear over potentially excessive collection of personal data, violating internationally recognized best practices for data collection and processing.
Understanding Facial Recognition Technology
Facial recognition is an algorithm-based technology capable of creating a digital map of a person’s face by identifying and mapping individual facial features. The technology then uses this digital map to search for matches in available databases.
In the Automated Facial Recognition System (AFRS), a large database of photographs and videos is utilized to identify individuals. An image of an unidentified person extracted from CCTV footage is compared with the existing database using artificial intelligence (AI) technology for pattern finding and matching.
The facial recognition system primarily works by capturing an individual’s facial features using a camera. Software reconstructs these features, storing the face and its features into a database. This database can be integrated with any software used for security purposes, banking services, and more.
Applications of Facial Recognition Technology
Facial recognition technology serves two primary applications – 1:1 verification and 1:n identification.
In 1:1 verification, a facial map is created to match against a person’s photograph in a database to authenticate their identity. This application is commonly used in modern smartphones to unlock devices.
1:n identification scans a photograph or video to match against a database, identifying the individual captured. Law enforcement agencies, including the Delhi Police, often use FRT for this application.
The need for such technology is attributed to its successful application in identification and authentication, with a success rate of nearly 75%. Especially in India, where there are only 144 constables per 100,000 citizens, this technology can act as a force multiplier, requiring little manpower and limited regular upgrades.
Why Delhi Police Uses Facial Recognition Technology
Delhi Police initially obtained FRT to trace and identify missing children, following the 2018 direction of the Delhi High Court in Sadhan Haldar vs NCT of Delhi. However, they later expanded its usage for police investigations, triggering concerns over ‘function creep’, where technology widens its scope from its original purpose.
The police have since used FRT during investigations into several high-profile incidents, including the 2020 northeast Delhi riots, the 2021 Red Fort violence, and the 2022 Jahangirpuri riots.
Potential Harms of Facial Recognition Technology
Despite its numerous applications, facial recognition technology is not without flaws. Inaccuracy and misuse top the list, with issues like misidentification due to inaccuracy and mass surveillance due to technological abuse.
Research also shows that FRT’s accuracy significantly decreases based on race and gender. This disparity can lead to false positives, where individuals are misidentified, or false negatives, where individuals aren’t recognized accurately.
Such errors can result in biases against misidentified individuals and exclusions for those unrecognized from accessing essential schemes. These errors have led to concerns over privacy violation and the absence of robust data protection laws.
The Way Forward
In the digital age, data is a highly valuable resource that requires strong regulation. India currently requires a robust data protection regime that respects citizens’ privacy while also enhancing the right to information.
Given India’s large population and understaffed administration, well-planned use of such nascent technology could be a solution, provided there are sufficient safeguards to address inherent concerns, including privacy issues.
