Every year during monsoons, police in western state of Gujarat in India organize Lord Jagannath Rath Yatra– a 16 kilometre long procession of roughly 1-1.5 million devotees who walk through the densely populated lanes of UNESCO World Heritage City Ahmedabad.

Here, hundreds of CCTV cameras become the eyes and ears of Gujarat Police as a centralized command and control centre keeps the cops on their toes, crowd-managing over a million devotees.

But what if the stakes are put higher?

80,000 CCTV cameras installed at 26 government departments spread across 34 districts of Gujarat. Can they be integrated into one centralized, effective system through artificial intelligence?

That is the challenge Gujarat Police is putting before India’s AI and deep-tech ecosystem through its Innovation Hackathon 2026 titled “Protect What Matters”.

The initiative is seeking production ready AI systems that can bring together feeds from different CCTV systems and turn hours of video into searchable, actionable intelligence for police operations.

There is a prize money pool of over 5 million Indian rupees for winning teams at the end of Hackathon. The last date to register and submit the proposals has been extended to October 12-13.

The overarching challenge is an “Integrated Video Management & Analytics Platform”, with a focus on AI-powered video analytics, real-time intelligence, person and vehicle identification, automatic number plate recognition (ANPR), event detection, tracking and Geographic Information System (GIS) based visualization.

The hackathon is looking for solutions that can integrate these systems, process live feeds at scale, apply AI-based analytics and correlate information to support faster decision-making. It does not focus on any particular AI model. Instead, participants are asked to propose practical and scalable architectures that can address the whole operational challenge.

The challenge comes at a time when CCTV networks are generating enormous amounts of footage, but the real difficulty lies in quickly finding and interpreting information hidden within that data. The state-wide CCTV system is large, heterogenous and fragmented with some relying on local storage while some on cloud storage.

Recently, Gujarat Director General of Police (DGP) Gyanendra Singh Malik asked police superintendents and commissioners in the state to encourage residential societies and commercial establishments to install CCTV cameras at isolated stretches and other vulnerable points in the cities and villages. Currently, privately owned 380,000 CCTV cameras are installed all across the state.

Gujarat Police wants participating teams to address that gap so that instead of relying on officers to manually monitor multiple screens or search through hours of recorded footage, AI can be used to find actionable intelligence and do live monitoring.

AI FrontPage reached out to Gujarat Police for a detailed interview on the hackathon. Published below are excerpts from the statements of official spokesperson for State Crime Records Bureau (SCRB), Gujarat Police.

Question: What specific public-safety challenges is Gujarat Police hoping to address through this hackathon?

Gujarat SCRB Spokesperson: The hackathon focuses on addressing challenges related to integrated CCTV management, video analytics, real-time intelligence, vehicle/person identification, event detection and actionable alerts. The objective is to develop solutions capable of bringing together feeds from diverse CCTV systems and converting video data into useful, searchable and actionable intelligence for public-safety operations.

Question: Why has AI-powered CCTV integration and real-time intelligence been identified as a key priority?

Gujarat SCRB Spokesperson: With a large and heterogeneous CCTV ecosystem, one of the key challenges is the ability to integrate different camera/VMS systems and analyze video feeds at scale. The challenge therefore focuses on creating a unified platform capable of processing live CCTV feeds, applying AI/video analytics, correlating information and generating real-time alerts to support faster and more informed decision-making.

Question: What kind of AI, video analytics or surveillance innovations is Gujarat Police particularly looking for from startups?

Gujarat SCRB Spokesperson: The hackathon welcomes innovative solutions around AI-powered video analytics, person and vehicle identification, ANPR, event detection, tracking, searchable video intelligence, real-time alerts, GIS-based visualization and integration of heterogeneous CCTV systems. The objective is not limited to a particular AI model; participants are encouraged to propose practical and scalable architectures that address the overall challenge.

Question: How will the department evaluate whether a proposed technology is suitable for real-world, large-scale deployment?

Gujarat SCRB Spokesperson: Evaluation is based on the solution’s functionality, architecture, integration capability, analytics, scalability and performance on real CCTV feeds. The problem statement specifically envisages testing with approximately 50 heterogeneous cameras, live feeds, designated vehicle tracking, alert generation/history and GIS visualization. Solutions are therefore expected to demonstrate operational capability rather than only a conceptual prototype.

Question: Can you share examples of existing technology or AI systems being used by Gujarat Police that startups could potentially build upon or integrate with?

Gujarat SCRB Spokesperson: At present, there are no specific existing technology or AI systems identified by the hackathon that startups are expected to directly integrate with. The challenge is focused on encouraging participants to develop their own solutions based on the requirements outlined in the problem statement.

Question: How does Gujarat Police plan to ensure that AI-based CCTV and video analytics systems are accurate and reliable in real-world conditions?

Gujarat SCRB Spokesperson: The challenge deliberately incorporates real-world and heterogeneous CCTV conditions into evaluation. Participants are expected to demonstrate their solutions against live/representative feeds rather than relying solely on controlled demonstrations. The evaluation considers the solution’s actual functionality, analytics capability, integration, scalability and performance under the provided CCTV conditions.

Question: What safeguards are being considered around privacy, data protection, algorithmic bias and responsible use of AI?

Gujarat SCRB Spokesperson: Any operational deployment of a successful solution would be subject to the applicable government policies, legal requirements, data-protection requirements and departmental protocols.

Question: What are the expectations regarding interoperability with existing CCTV and command-and-control infrastructure?

Gujarat SCRB Spokesperson: Interoperability is a central part of the challenge. The proposed architecture may involve integration with departmental CCTV/VMS systems through technologies such as RTSP, ONVIF, vendor SDKs and available APIs, along with suitable streaming/relay mechanisms such as HLS or WebRTC. Participants can choose among different integration models, including unified viewing, VMS federation/middleware, central VMS and hybrid/innovative architectures.

Question: Could successful solutions eventually be deployed across multiple cities or districts in Gujarat?

Gujarat SCRB Spokesperson: The challenge is designed with scalability and real-world deployment in mind. The evaluation itself considers operation across a large heterogeneous camera environment. However, any deployment across cities or districts would depend on the solution’s technical suitability, operational requirements and the applicable government approval and procurement processes.

Question: What would success look like six or twelve months after the hackathon?

Gujarat SCRB Spokesperson: From the department’s perspective, success would mean identifying solutions that can progress beyond a hackathon demonstration and demonstrate technical maturity, interoperability, scalability and practical applicability to policing and public safety. The longer-term objective is to explore how promising technologies can contribute to real-world police operations.

Question: What message would you like to give to Indian AI and deep-tech startups considering participation?

Gujarat SCRB Spokesperson: We encourage startups, developers, innovators and technology companies to view this not simply as a competition, but as an opportunity to solve a genuine public-safety challenge using technology. The challenge provides access to real CCTV-based testing conditions and asks participants to think beyond a basic prototype—to build solutions that are scalable, interoperable, reliable and capable of addressing real operational requirements.

Also Read: 2,000 Cameras, A Million Violations, One Algorithm: How AI Took Over Chandigarh’s Roads