Ever felt like someone was watching your every move even when you were doing something as innocuous as a grocery-run? For the most part, it’s just your imagination playing tricks on you. However, such a form of state surveillance may soon become a reality in Chennai where city police plans to track and identify citizens in a bid to reduce crime.
The Greater Chennai Police (GCP) has floated a tender to set up an automatic number-plate recognition (ANPR) system at 50 locations within its jurisdiction. As per the document, the system will help police officers solve vehicle thefts and track criminals.
However, what it glosses over in its text are the technical details that indicate the police will be maintaining a “face database” or “face library” of all the people that are detected by the AI-based cameras. Worse still, the specifications of the system asks the technology to track and record the route that every face takes – essentially monitoring the movement of the residents without their permission.
STAY ON TOP OF TECH POLICY: Our daily newsletter with top stories from MediaNama and around the world, delivered to your inbox before 9 AM. Click here to sign up today!
Why it matters: Granted that with MediaNama’s frequent coverage of the GCP using facial recognition systems (FRS), the idea of a face library may not appear new. However, this time the police is talking about ‘tracking faces’ as well which affects not one but two fundamental rights of privacy under Article 21 and freedom of movement under Article 19. It pulls closer the possibility of a surveillance state and the default viewing of innocent citizens as criminal suspects. This considered with the Criminal Procedure (Identification) Act, 2022 – CPIA and the Digital Personal Data Protection (DPDP) Bill, 2022 can have devastating effects on people’s rights.
Tracking people to save vehicles
Included in the deliverables of the system is the automatic tracking of the route followed by the stolen vehicle. For this the police made several data centre related specifications that included the following:
People Counting – Through this, the ANPR system will provide daily, weekly, monthly and yearly reports of people gathering or passing through specific regions. The counting will involve heatmapping and the system will export the people counting and heatmap data when required.
Face Database – This involves creating and managing a ‘face library’ to which a new face is added every time it is detected by the ANPR system. The database will import faces in batches and set a “person type for face.” This can also be done manually. The face library will be accessible to FRS devices to configure the similarity.
Face Recognition – The ANPR system automatically captures faces in the camera’s field of view and extracts face attribute information. The FRS then does a real-time face comparison and registers the face to the face library · After this, the facial recognition can search for faces in the library using attributes or an uploaded image of a face. It can even search “comparison records” as mentioned in the document.
Further specifications ask the system to “generate the track for the specific face” and to create a daily, weekly, monthly attribute report based on the gender and age.
“The application should be able to track the route of the identified stolen vehicle by integrating and analyzing real time data obtained from all other ANPR cameras,” said the document, meaning that the police will be able to track a target across cameras.
The police also asked for such tracking during live view. Accordingly, the live description also mentioned fisheye and speed dome smart tracking as well as smart tracking for panoramic camera.
Understanding the intersection between surveillance and existing laws
In April last year, the Lok Sabha passed the CPIA (then Bill) with a voice vote. Opposing MPs criticised this Bill not the least because it blurred the distinction between a suspect, undertrial and convicted. Moreover, it allowed the police to collect the biometrics of suspects and store the data for as long as 75 years – longer than the average lifespan of an Indian. The way this ANPR system will be set up, it is likely that many people may have their data processed for being “suspects” without the data principal’s knowledge.
Similarly, the DPDP Bill was supposed to provide people the safeguard to protect their personal data. However, with the introduction of deemed consent and the inclusion of ‘law and order’ the Bill allows the police to wantonly process people’s data regardless of their consent. A bigger concern is that it will allow the department to share the data with other government departments, violating the basic principles of privacy.
In December, the GCP already showed its lack of regard for these principles when it confirmed on Twitter that officials were forcibly scanning people’s faces at night to prevent crime.
Dubious attempts to protect people’s privacy
Who benefits with CCTV sign posts? In terms of privacy, the GCP asked that the locations with these AI-based CCTV cameras have “some standardized signs” informing people of the camera’s existence.
“This will bring about the transparency on installation of CCTV cameras and no one would be able to later complaint for breach of privacy,” said the document.
Yet, how will a sign help citizens refuse their consent to be recorded? Putting up a sign simply absolves the police from their responsibility to inform a data principal of the processing of their personal data. The tender doesn’t mention or discuss possible recourse for people who may want to withhold or withdraw their consent.
What’s the point of privacy masks? Another measure mentioned in the ANPR specification is a ‘privacy mask.’ This is like the blur feature in video calls that allows the background to be distorted to hidden completely. However, the data centre already talks about detecting the face of any person who appears on screen. So, what is the point of the privacy mask if the system has already declared it will process the biometric data of citizens?
GCP’s immediate plan of action: The police plans to set-up the ANPR system on all the selected locations. For this, it will first implement software by integrating CCTV cameras, Vahan – a state service that takes care of vehicle registration – and other applications to provide alert systems to the department.
After that, the administration will establish local data storage and processing, recognition of vehicles, alerts to zone offices, command centres and Police tabs. Coupled with the pre-existing usage of FRS, CCTV network and talks of drone police units in the future, Chennai may soon be known as the safest or the most monitored city in India.
This post is released under a CC-BY-SA 4.0 license. Please feel free to republish on your site, with attribution and a link. Adaptation and rewriting, though allowed, should be true to the original.
- Explainer: Learn About Chennai’s Mega City Policing A.K.A. ‘Surveillance’ Project
- Why Tamil Nadu Govt Must Reconsider Its Move Mandating Aadhaar For Schemes: Internet Freedom Foundation
- How To Explain Tamil Nadu’s 360-Degree Database To Your Peers
- Lucknow Safe City Project: UP Government Wants A Data Framework To Establish ‘Relationship Between Entities’