On 14 May 2022, Chief Minister M.K. Stalin launched an AI-enabled panic button and CCTV surveillance system across 500 Chennai Metropolitan Transport Corporation buses under the Nirbhaya Safe City Project. By 10 January 2023,eight months later,Youth Welfare and Sports Development Minister Udhayanidhi Stalin inaugurated a dedicated Integrated Command and Control Centre at MTC headquarters on Pallavan Salai. By that date, the system had expanded to 2,330 buses. A total of 6,990 cameras, 9,320 panic buttons, and 66 depot and terminus locations had been brought under surveillance. The project cost stood at Rs 72.25 crore, financed through the Nirbhaya Fund.
Between the Phase 1 launch and the Phase 2 inauguration, the Metropolitan Transport Corporation published no evaluation of the first 500-bus deployment. No alert rate data. No false-alert analysis. No data retention policy. No grievance mechanism specific to the surveillance architecture. The system moved from pilot to permanent public infrastructure in eight months without a visible accountability trail. That transition is the governance question.
The Architecture
The MTC system is not a conventional CCTV installation. Each covered bus carries three IP-based cameras, four panic buttons, a GPS tracker, and one AI-enabled Mobile Network Video Recorder. When a passenger presses a panic button, the MNVR captures 60 seconds of footage and triggers an alarm at the ICCC. Sixteen operators monitor a 40×7-feet video wall at the ICCC around the clock. Alerts are relayed in real time to the Greater Chennai Police distress response centre.
The technical architecture carries governance implications that the contracts have not yet publicly resolved. MTC’s Request for Proposal (RFP No. 17/44017/Proj/MTC/2020) specifies that the system integrator may, in the future, be required to transmit video feeds to “any other CCC run by government agencies (like Chennai Smart City, GCC, GCP, etc.).” This inter-agency interoperability clause appears inside a procurement document. It has not been publicly mapped onto a disclosed data-sharing protocol between the receiving agencies, nor has any retention standard been published for each transmission endpoint.
The system was also designed with a scope wider than its stated mandate. The government’s own May 2022 announcement noted that the project would “aid in detecting missing persons and identifying criminals and other works of the GCC, transport department and the police”. This is a different function from responding to panic alerts. The two objectives co-exist in the same infrastructure without a disclosed framework for distinguishing when each applies, who authorises the broader use, and under what procedural safeguards.
The Accountability Gap
At the ICCC inauguration, an MTC official stated that “so far no incidents affecting the safety of women and children happened in the bus that required pressing of the panic buttons”.The observation was presented as reassurance. It is also a disclosure of baseline absence: after eight months of operations across 500 buses, no public performance report was released before the rollout was expanded nearly fivefold.
Institutional expansion and institutional accountability did not move at the same speed.
Tamil Nadu’s Safe and Ethical AI Policy, released on 19 September 2020, applies explicitly to all Tamil Nadu government bodies, PSUs, and organisations controlled by or receiving aid from the state government (TN Safe & Ethical AI Policy, 2020). The Metropolitan Transport Corporation is a state PSU. The policy requires procuring departments to consult the Centre of Excellence in Emerging Technologies at TNeGA before any AI acquisition. It mandates a DEEP-MAX scorecard evaluation across seven parameters—Diversity, Equity, Ethics, Privacy and Data, Misuse Prevention, Audit and Transparency, and Cross-Geography performance, before any AI system is deployed for public use. It requires periodic updates to the DEEP-MAX score every six months post-deployment and specifies that scores must be logged on an ACTS-Blockchain maintained by TNeGA.
No publicly available record confirms that the MTC’s AI-enabled MNVR and panic button system underwent a DEEP-MAX evaluation before Phase 1 or Phase 2 deployment. The Safe and Ethical AI Monitoring Committee, headed by the Chief Secretary, is the designated oversight body under the policy. Its deliberations on this deployment, if they occurred, have not been published. Whether the CEET at TNeGA reviewed this system before procurement has not been answered in any publicly accessible document.
The gap between the policy’s governance architecture and the MTC deployment’s operational trail is not incidental. The DEEP-MAX parameters; particularly Privacy and Data, Misuse Prevention, and Audit and Transparency are precisely the standards that a continuous surveillance system processing footage of public commuters is required to satisfy before going live.
The Data Protection Question
The Digital Personal Data Protection Act, 2023 is India’s operative data protection legislation. Section 17(2)(a) of the Act provides that its provisions shall not apply to the processing of personal data by any instrumentality of the state that the Central Government may notify, in the interests of sovereignty, integrity, security of the state, friendly relations with foreign states, maintenance of public order, or preventing incitement to any cognizable offence (MeitY, DPDP Act, 2023). The Carnegie Endowment for International Peace has noted that “state surveillance agencies have been consistently exempted from the application of data protection requirements” under this Act (Carnegie Endowment, October 2023).
MTC, as a state public sector undertaking, may fall within this exemption category. That has direct practical consequences. The MTC surveillance system processes visual footage of every passenger who boards a covered bus across 2,330 vehicles and 66 terminal and depot locations. No publicly disclosed policy specifies how long this footage is retained. No publicly disclosed policy specifies the conditions under which police may access it, how inter-agency sharing will be governed once feeds are transmitted to Greater Chennai Police or Chennai Smart City, or by what process a passenger whose image has been mis-identified or shared in error can challenge that record.
The DPDP Act’s notice, consent, and erasure provisions do not structurally bind a state entity that may be notified as exempt. The RFP’s inter-agency transmission clause creates a chain of data movement with no publicly articulated retention standard or access protocol at each receiving end. The absence of these protocols is not a legal technicality. MTC buses are used disproportionately by low-income, working-class, and women commuters. The populations most likely to be surveilled are also least likely to hold institutional leverage to contest how their footage is used.
The National Pattern
Chennai is not exceptional. It is a point in a recognisable national pattern.
The Nirbhaya Safe City Project was approved in 2018 for eight cities with a combined outlay of Rs 2,919.55 crore: Delhi (Rs 663.67 crore), Bengaluru (Rs 667 crore), Mumbai (Rs 252 crore), Kolkata (Rs 181.32 crore), Hyderabad (Rs 282.50 crore), Ahmedabad (Rs 253 crore), Lucknow (Rs 195 crore), and Chennai (Rs 425.06 crore) (PIB, August 2018). Each city received central funds to build surveillance infrastructure under a women’s safety mandate. In each city, the mandate preceded the governance framework.
A Comptroller and Auditor General audit of the Nirbhaya Fund in Gujarat, published in August 2024, found Rs 143.75 crore locked in committed liabilities and Rs 57.66 lakh unutilised under the Safe City Project as of March 2023, “even after a period of three to five years.” The CAG report stated: “The Home Department needs to evaluate the progress of the scheme”. Whether a comparable public evaluation has been conducted in Tamil Nadu has not been disclosed.
Under the Smart Cities Mission, more than 84,000 CCTV surveillance cameras have been installed across 100 Indian cities (Rajya Sabha Unstarred Question No. 1490, August 2024). The India AI Governance Guidelines released by MeitY in February 2026 call for mandatory grievance redressal mechanisms, transparency reports, and accountability frameworks for AI deployments (PIB, February 2026). These guidelines are voluntary. They carry no binding obligation on state government surveillance systems already in operation.
Tamil Nadu’s own AI Mission, established under G.O. Ms. No. 25 in 2024 with a two-year allocation of Rs 13.93 crore, lists “security and public safety” as an explicit AI use case domain (TN Government Order, 2024). The mission document does not reference how existing AI surveillance deployments including the operational MTC system, will be retroactively assessed under its emerging framework.
The Open Question
The MTC system is permanent infrastructure. It is not a pilot pending assessment. It processes footage of every passenger boarding an equipped bus across a network that moves millions of commuters every week. Its command centre operates around the clock. Its data feeds are architecturally designed for inter-agency transmission.
The system’s stated objective safety for women in public transport is a legitimate and funded public goal. The question the state has not yet answered publicly is narrower than that goal: by what mechanism does a commuter whose image has been retained, mis-flagged, or transmitted between agencies know that this has occurred, and by what process can she contest it?
Tamil Nadu’s Safe and Ethical AI Policy sets out the relevant evaluation standards. MeitY’s AI Governance Guidelines identify the relevant institutional mechanisms at the national level. Neither has been demonstrably applied to a deployment that was operationally live across more than two thousand buses before either framework had a visible role in governing it. A Right to Information application to TNeGA asking whether a DEEP-MAX score was filed for the MTC system, and to MTC asking for its data retention policy, would begin to establish the record. Until that record is public, what Chennai has built is a well-funded surveillance system with a safety mandate, a governance policy on paper, and an accountability trail that has not been opened to scrutiny.