The use of AI by law enforcement agencies highlight the need to evaluate governance considerations within this ecosystem.
Image Source: Getty Images
Law Enforcement Agencies (LEAs) are increasingly using Artificial Intelligence (AI) to enhance their functioning, particularly with enhanced capabilities for predictive policing.
Globally, there is an increase in the adoption of new technologies by law enforcement. In the United States (US), the New York Police Department has adopted tools like Patternizr for pattern analysis and officer deployment. Similarly, in China, the government uses robots for crowd control and monitors suspicious activities using drones and detention cameras. Scientists are developing a virtual reality model of Shanghai, including offices and family residences to provide real-time assistance to police and rescue services. Both the US and Australia have also focused on child protection using AI. Initiatives like America’s use of Clearview AI and the Australian Centre to Counter Child Exploitation enable faster threat detection and prevention in child exploitation cases. Meanwhile, South Korea has introducedAI patrolling cars , which integrate voice recognition, video analysis, and real-time data processing to ensure security on roads. These examples highlight a worldwide trend of integrating tech tto address policing challenges.
Scientists are developing a virtual reality model of Shanghai, including offices and family residences to provide real-time assistance to police and rescue services.
The global market size for predictive policing alone is estimated to rise to US$157 billion by 2034, with a CAGR of 46.7 percent during 2025-34. The prospect of integrating huge criminal datasets for an expedited investigation process has been of great appeal to governments, including India. The police-to-population ratio in India is 153 per 100,000 people. This is below the 222 per 100,000 people advised by the United Nations. The reduction of this gap and increasing resource distribution efficiency are some of the motives to include technology in law enforcement.
The applications of AI-driven law enforcement range from counterterrorism to crowd management. In Uttar Pradesh use, the use of AI-powered drones and CCTVs came handy for tracking people and managing traffic in large gatherings like the Kumbh Mela. This indicates a broader adoption of technology, mirrored by unique applications such as preserving and digitising fingerprints with greater accuracy in criminal forensics. Moreover, modern tools developed by central agencies like the Bureau of Police Research & Development (BPR&D) delve into spaces like the deep and dark web to gauge sentiment and provide credible intelligence input to LEAs.
India has also attempted to curb new-age cybercrimes like online money laundering. The Enforcement Directorate uses advanced analytical AI/ML tools of the Financial Intelligence Unit (FIU) to detect suspicious monetary patterns. The FIU’s data analysis of mule accounts also helps LEAs in preventing routing of unaccounted money in the form of Virtual Digital Assets.
However, AI systems sometimes get overwhelmed and provide suboptimal performance. Despite having 275 AI CCTVs, the Rath Yatra in Puri witnessed the demise of three pilgrims. Technical inconsistencies like false positives—targeting people with darker skin tones, as documented in countries like the US—due to algorithmic biases, raise additional concerns in a diverse nation like India. Such instances raise crucial questions regarding the accountability of technology service providers and governance considerations.
The endeavours of the states to equip and modernise their police forces have been augmented by the centre via the “Assistance to States & UTs for Modernisation of Police” (ASUMP)—a INR 4,846 crore outlay for a period of five years between 2021-26. Delhi and Tamil Nadu are said to have employed ‘Innsight’—an AI tool for data analysis, developed by Innefu Labs, a private firm which has been subject to cyberattacks and data breaches on account of a weak security structure. Cases like these suggest the need for a framework of due diligence for private firms in securing contracts, reinforcing the need for due process and adequate testing before deployment. While AI tools promise efficiency, their deployment needs to be accompanied by mechanisms for explainability to ensure requisite feedback loops and accountability to counter occurrences of opaque behaviour.
A governance framework for the deployment of AI in law enforcement would need to account for the potential for bias, discrimination, and false positives, raising questions of liability and accountability. This requires reconciling operational use with passing tests of legality, necessity and proportionality, particularly within the context of the fragmented regulatory landscape for biometric data. Without such a framework, it will be difficult for the government to gain the trust of the public, especially in the age of social profiling—where every action on the internet is of interest to multiple surveilling entities.
LEAs can leverage this to perform proactive policing by scanning large amounts of data quickly, build better preventive mechanisms for new-age crimes like cyberattacks, and address resource allocation efficiently.
However, safeguards and regulatory frameworks need to be complemented with skill and capability on the ground. The case of the use of generative AI to write police reports in the US has highlighted how the AI-generated reports missed the contextual specificities of particular jurisdictions and legal nuances of policing practice.
India’s current legal and operational frameworks must carefully incorporate protections and standards to close these institutional and technical divides.
The integration of technology in law enforcement has immense potential to increase the productivity of officers and streamline the process of apprehending offenders. LEAs can leverage this to perform proactive policing by scanning large amounts of data quickly, build better preventive mechanisms for new-age crimes like cyberattacks, and address resource allocation efficiently. However, the responsibility for any mishap during the deployment of technologies in law enforcement also rests primarily with the operating agency in India.
To keep up with the pace of technological growth and use AI responsibly, LEAs must embrace evaluation mechanisms to accompany technological evolution.
Companies should undergo periodic external algorithmic audits and receive a compliance certification from the auditing authority. Additionally, clear compliance mechanisms must be conveyed to companies to be eligible for procurement by LEAs. The role of an Artificial Intelligence Safety Institute in developing robust safety and ethical testing standards, suitable for an Indian context, can be a key policy initiative to execute further.
Pilot programmes should be made mandatory to determine the real impact of the AI and evaluate the same against risk parameters within a given context. The proposition to establish an ‘Incident Database’ that will create a collection of risks and build harm reduction mechanisms related to the deployment of AI tools will help in recognising the evolving nature of harms. All of these proposed mechanisms will require deep collaboration between the public and private bodies, and especially with the developers of these new technologies.
The role of an Artificial Intelligence Safety Institute in developing robust safety and ethical testing standards, suitable for an Indian context, can be a key policy initiative to execute further.
Moreover, there should be sensitisation efforts within LEAs to train their officials in the field on the responsible use of AI, including knowledge of potential risks and harms. Some AI monitoring systems can detect inconsistencies in police activities by identifying patterns of prejudice and disproportionate force among police officers, helping in internal performance evaluations. The BPR&D can set up dedicated AI and technical training modules across police verticals and play a crucial role in this context.
Thoughtful regulation, along with human oversight, is the cornerstone of developing effective governance frameworks to ensure the development of safe and trusted systems for a diverse nation like India.
Srijan Jha is a Research Intern at the Observer Research Foundation
The views expressed above belong to the author(s). ORF research and analyses now available on Telegram! Click here to access our curated content — blogs, longforms and interviews.
Srijan Jha is a Research Intern with the Centre for Security, Strategy and Technology at the Observer Research Foundation. ...
Read More +
PREV