Smart FIRs - Leveraging AI to Improve Efficiency and Trust in India’s Police System
Authors
Executive Summary
Megha Nambiar serves as the General Counsel at HyperVerge, where she leads fintech-focused legal strategy, regulatory compliance, and business expansion. She is an alumna of the Takshashila Institution.
Kartik Singh is an ex-Staff Software Engineer at Google for 18 years. He is currently a PGP student at the Takshashila Institution.
This policy brief focuses on reimagining the First Information Report (FIR) registration process, suggesting its augmentation using the capabilities of artificial intelligence (AI) – thereby making it faster, more accurate and less tedious.
Alternatives have been evaluated on two parameters - levels of technology incorporated in the process, and identification of the right stakeholder(s) for efficient implementation. The selection criteria envisaged are a) equity, accessibility and inclusion, b) FIR velocity (total time taken to complete the FIR process), c) cost and operational efficiency, d) user trust and accountability, and e) accuracy and completeness. Privacy, data protection and compliance with law serve as essential boundary conditions that need to be met by all alternatives.
This policy brief identifies the most suitable approach and proposes a four-phase pilot run over 17 months in two districts (one urban and one rural) in the State of Karnataka. The brief stipulates clear goals on the basis of which the pilot’s impact should be measured, and if successful, be scaled nation-wide offering a path to a more efficient, convenient and citizen-oriented process for justice delivery.
1. Introduction
First Information Reports (FIRs) are central to the criminal justice process, yet their registration is often inconsistent, time-consuming and inaccessible, undermining overall police efficiency and public trust. Existing modernisation of the system and adoption of digital best practices is varied across states, especially vis-à-vis the adoption of Crime and Criminal Tracking Network and Systems ( CCTNS), e-Zero FIR, Inter-Operable Criminal Justice System (ICJS) integration, and state eFIR Portals. Several challenges, such as incorrect data collection, lack of appropriate use of technology, improper synchronisation of manual and digital processes that run in parallel, etc., exist. 1
The visual below summarises the chain of events for the FIR registration process in India.
Summary of the Key Steps in the FIR Registration Process in India
2. Systemic Challenges
This section explores the current systemic gaps and areas where policy intervention, specifically through technological integration, can ameliorate the current challenges.
| Challenge | Description |
|---|---|
| Victim intimidation | There exist several barriers discouraging victims from registering FIRs. These may stem from cultural norms, fear of police intimidation, or distrust in the process. In Lalita Kumari v. Govt. of U.P. & Ors 2, it was observed that “Keeping in view the NCRB figures that show that about 60 lakh cognizable offences were registered in India during the year 2012, the burking of crime may itself be in the range of about 60 lakh every year.” Similarly, a survey in Maharashtra found that FIRs are not filed for 50 per cent of the cases. 3 |
| Manual paperwork | FIR registration relies heavily on manual paperwork, which hampers systemic data analysis for crime prevention efforts. Information recorded is prone to errors, omissions and discretion. Currently, India only has one computer/laptop for 11 police personnel. 4 |
| Understaffed police force | The administrative burden of FIR registration leads to inefficient utilisation of already limited police resources and personnel. The ratio is ~156 officers on average per 1 lakh persons, and the ratio is 81 in states like Bihar, contrasted to the highly digitally empowered police force of the US which still retains over 210 officers per 1 lakh persons. 5 6 7 |
| Long wait times and delays | FIR registration often involves significant waiting periods and delays. In addition to the aforecited judgement of Lalita Kumari v. Govt. of U.P. & Ors, in Sangita Devi v. Govt of U.P. & Ors, the police disregarded a court order for four months before registering an FIR.8 |
| Limited transparency and accountability | The transparency and accountability within police stations remain very limited, as evident from the examples cited above. |
| Fragmented digital infrastructure | Given that ‘Police’ is a state subject, the implementation process followed for FIR registration is inconsistent across districts and states, creating barriers in synchronising the data across the nation.9 |
| State of tech adoption | As per the CAG report on Assam, synchronisation of FIRs into the central database is delayed by 170 days on average, when it should occur within 24-72 hours. Further, data points in FIRs in the CCTNS have seen significant inaccuracies when compared with the data points filled by the registering police officer.10 |
| Low accuracy | Due to human input errors, issues occur during the process of recording FIRs by hand and subsequently typing them in digital form, increasing the possibility of inaccuracies.11 |
3. Levels of Technology
This section outlines strategies to address these issues, focusing on strengthening the FIR registration process through the adoption of modern technologies and streamlined procedures.
3.1. Status-Quo
The present process is largely manual – the informant visits a police station, the officer puts the oral information into writing, reads it back, and the informant authenticates it through a physical signature or thumb impression. Thereafter, a numbered FIR is issued, entered in station records, and forwarded The present process is largely manual – the informant visits a police station, the officer puts the oral information into writing, reads it back, and the informant authenticates it through a physical signature or thumb impression. Thereafter, a numbered FIR is issued, entered in station records, and forwarded to the Magistrate and senior officers. The current process is labour-intensive, contains elements of discretion, is lower on accessibility and coverage, and time-consuming. Not to mention,the utilisation of technology for case management/tracking varies across states, and even when technology is implemented, its adoption is disparate.
3.2. Virtual Approach
This approach maximises use of technology, reducing human manual intervention in the FIR registration process. The end-to-end FIR filing process shall be carried out over a mobile application or web platform. The informant’s statement shall be recorded for future analysis in the form of the video/audio. An AI conversational assistant (Interactive Voice Response (IVR)) shall guide the informant and provide all the relevant details to file the FIR correctly, asking context-based questions and collecting all necessary information. Speech-to-text shall convert oral statements into a digital FIR copy, which shall be displayed and also read back to ensure the informant is fully aware of the contents of their FIR, and confirms the same before it is registered. Real-time multi-lingual translation of text can be applied where necessary. A copy of the same shall be provided to the informant as required under Bharatiya Nagarik Suraksha Sanhita.12
For authentication, the current practice often requires the informant to physically visit the police station to physically sign the FIR. Instead, e-KYC mechanisms and Aadhaar-based e-signatures can be utilised, in accordance with Section 173 of BNSS. Artificial Intelligence can assist in identifying or recommending specific provisions of Bharatiya Nyaya Sanhita/charges which should be applied in the FIR, to be reviewed and confirmed by the respective police officer. A centralised dashboard shall track each FIR against Service Level Agreements (SLAs) accessible by officers and informants, improving transparency and accountability. Routing of the FIR to the correct police station will be carried out through technology.
A self-serve kiosk shall be installed within police station premises and Common Service Centres (CSCs) to serve individuals without digital devices.13 In case of any technical issues or if the informant is unable to use the technology, they will be able to raise a ticket for support, which will be addressed by a chatbot or escalated to support personnel as appropriate.
3.3. Hybrid Approach
This hybrid model integrates key elements of the virtual process outlined in Approach 2 with the existing procedure for filing FIRs, thereby accommodating both in-person and virtual registration. The AI assistant will serve as a valuable tool for both the informant and the police officer responsible for recording the FIR. This system will assist officers in formulating follow-up questions based on initial information and will generate an analysis to streamline the process. This system differs from Approach 2 in that it will not operate end-to-end autonomously, but instead the efficiencies of technology will be combined with the experience and judgement of police personnel, who will be aided by the prowess of these systems.
This approach prioritises informant autonomy by ensuring they can record their statement without alteration or manipulation. The officer’s role will be to ensure all necessary details for a comprehensive FIR are collected. Like Approach 2, this model will incorporate e-KYC, Aadhaar based e-signatures and case management dashboard elements to enhance transparency and ensure SLA compliance. The benefits of speech-to-text, real-time translation if required, video/audio recording, and valid online authentication signatures will continue to apply in this Approach. The AI assistant will also facilitate a final review of the FIR before official submission. Assisted-digital kiosks and Common Service Centres (CSCs) will be available for citizens without digital devices.
Ultimately, this hybrid approach brings together the substantial benefits of technology with a crucial human-in-the-loop component. This design prioritises review and safety, ensuring that complex and sensitive matters benefit from the human touch, holistic judgment, and discretion required for effective and just outcomes.
4. Implementation Stakeholder
There are three ways in which the integration of technology into policing can be implemented by: a) complete government ownership; b) public private partnership, ensuring government oversight over the private entity which will be involved to incorporate technology; and c) full privatisation with the entire process handed over to a third-party.
In the first case, the government will follow the “do it yourself” model and implement the process from end-to-end. The government will be responsible for the entire process including determining the technology, deployment of the infrastructure, training the police officers on the new process etc.
The second case involves privatising and outsourcing elements of the FIR filing process to third parties under a public-private partnership with the oversight of the police officers. The private company would assume responsibility and determine the technology to be deployed subject to defined boundary conditions in line with the requirements of the state. For example, emergency call handling in other countries have similarly been outsourced entirely to private entities as seen in the ChatComm model.14 When the end-to-end first information collection is outsourced, police officers will be involved in the final review, verifying provisions and authorising the filing of the FIR freeing up their time for higher order activities such as investigation. SLA compliance will be enabled using stringent KPIs for the third-party entity involved. In India, a similar approach is also used in the Common Service Centres (CSC) model for government services.15
In the last case, the complete responsibility of the FIR registration process will be transferred to the chosen third-party. The third-party will be able to determine the technology, invest in the infrastructure to deploy the solution, hire the human resources required and conduct required training within defined parameters. The police will only step in once the FIR has been registered.
5. Analysis
5.1. Analysis of Levels of Incorporated Technology
To carry out the analysis of the levels of technology to be incorporated into the FIR registration process, the following criteria have been used:
| Criteria | Description | Measure |
|---|---|---|
| Equity, Accessibility and Inclusion | The system must be equitable, accessible to all, unbiased, while maintaining the accuracy and integrity of information. | The number of languages covered, accuracy and fidelity levels of text in FIR, and demographic access parity. |
| FIR Velocity | Optimisation and systemic consistency in the time from the initiation of the complaint to the final issuance of the FIR. | Timestamps linked to pre-defined SLAs. |
| Operational & Cost Efficiency | The total cost, including software, hardware, training, installation, and maintenance per FIR. | Operational efficiency of the time spent by an officer per FIR. |
| User Trust & Accountability | The level of confidence that the system provides complainants that their complaint has been received, logged, and will be acted upon, combined with the effectiveness of follow-up mechanisms for grievances. | User survey (Net Promoter Score), analysis of ticket closure for grievances and turnaround times for completion of process. |
| Accuracy & Completeness | The extent to which the process ensures coverage of all mandatory statutory fields, error rates of text recorded based on the informant’s information and the precision of legal provisions applied. | Random sampling audits and analysis of % of FIRs that are not quashed or corrected later due to filing errors. |
| Privacy and data protection | Ensure the solution fully upholds privacy and data protection imbibing secure storage with appropriate access controls and robust information security infrastructure. | Privacy and security audits, frequency of information security incidents. |
| Compliance with current law | Adherence to Bharatiya Nagarik Suraksha Sanhita, Bharatiya Sakshya Adhinayam, and the Digital Personal Data Protection Act for admissibility in judicial proceedings. | Adherence to the chain of custody requirements, compliance with data protection laws and verification and fidelity of the process followed. |
Equity, Accessibility and Inclusion: Technology plays a vital role in broadening the reach and improving the accessibility of public services. The Indian populace has demonstrated a rapid embrace of new technologies, as evidenced by the widespread adoption of digital platforms like the Unified Payments Interface (UPI). The status quo presents a notable barrier to equity and inclusion. In the existing manual process, a complainant who is not fluent in the local language in which the FIR is being recorded is often required to sign a document they cannot read or fully comprehend. In contrast, the Virtual and Hybrid approaches can facilitate real-time translation and provide the complainant with a version of the FIR in their preferred language, ensuring they understand the contents before affixing their signature.
FIR Velocity: The status quo is inefficient, as it requires a police officer to manually transcribe the details of the FIR, which is a time-consuming process. In contrast, the Virtual approach significantly accelerates FIR registration by leveraging automated tools such as speech-to-text technology and AI assistants. These technologies can guide the complainant and instantly convert spoken information into text, thereby streamlining the entire process. The Hybrid approach offers a balance between speed and quality. The officer’s supervision and final approval will inevitably introduce a slight delay compared to a fully automated system.
Operational and Cost Efficiency: The implementation of either the virtual or hybrid approach for FIR registration will necessitate a substantial one-time capital investment in technological infrastructure. This expenditure, however, will yield long-term operational efficiency and cost savings by reducing the reliance on human resources for manual tasks. Conversely, while maintaining the status quo avoids immediate technology-related costs, it incurs a significant and continuous expense due to the inefficient allocation of valuable police personnel.
User Trust and Accountability: Both the Virtual and Hybrid approaches for FIR registration offer significant advancements in system transparency and accountability. By enabling robust tracking capabilities, ensuring adherence to SLAs, and integrating a comprehensive grievance mechanism, these systems will improve public trust. However, while the Virtual approach increases transparency through digitisation, it lacks human oversight, which may erode user confidence.
Accuracy and Completeness: The current status quo relies entirely on a police officer’s manual data entry, which can lead to inconsistencies and errors in accuracy and completeness, particularly when complaints require translation. In the Virtual approach, the accuracy of the FIR would be contingent on the reliability of AI technologies, specifically the quality of speech-to-text conversion for various regional languages and accents. It would also depend on the AI assistant’s ability to effectively guide a complainant to provide a complete report. By contrast, the Hybrid approach combines human judgment with AI assistance, resulting in significantly higher-quality FIRs. This model uses AI to improve data capture and content, while the human element verifies the accuracy and contextual relevance of the final document.
Privacy and Data Protection: The centralisation of data, which is an inherent feature of both the Virtual and Hybrid approaches, introduces significant cyber security risks, including the potential for unauthorised access and data breaches. Given that the full digitisation of police records is an eventual reality, with many police stations already at varying stages of digital integration, it is critical to address these vulnerabilities proactively and enforce stringent data security measures and user privacy protocols from the outset.
Compliance with Current Law: Compliance with the Hybrid approach is largely achievable and even supported by the newly passed BNS, BNSS and BSA, which all veer in the direction of higher technology adoption in the existing processes. In the context of the Virtual approach, which removes officers from the process, the same does not find legitimacy in the current legal framework, which very much requires the intervention of police officers in the FIR process, making the Hybrid model a more pragmatic and immediately achievable path forward.
The following image shows the comparison across the criteria and essential boundary conditions of the three levels of incorporated technology:
Comparison of Criteria and Essential Boundary Conditions of the Three Levels of Incorporated Technology
Based on the analysis, the Hybrid approach offers significant advantages over the current system and over a fully automated process for registering FIRs in India. It leverages the strengths of both technology and human expertise to build a more robust and trustworthy law enforcement process.
5.2. Analysis of the Implementation Stakeholder
To carry out the analysis of the composition of the implementation stakeholder, the following criteria have been used:
| Criteria | Description | Measure |
|---|---|---|
| Scalability & Technological Expertise | The implementer must possess the technical expertise and infrastructure to deploy and manage the solution effectively across an entire state, with the potential for nationwide expansion. | The implementer’s proven track record, the existing technical knowledge and the scale of their previously executed projects. |
| Cost Efficiency | The chosen entity should be structured to ensure the solution is delivered in a cost-effective manner. The system must incentivise operational efficiency and discourage inflated expenses. | Performance indicators related to employee productivity and resource utilisation, benchmarked against industry standards. |
| Legal Authority | The implementing body must be legally empowered to execute all aspects of the solution, including the legally binding process of First Information Report (FIR) registration. | Legally recognised authority, as established through official mandates or legislation, to oversee and operate the FIR registration process. |
| User Trust & Public Legitimacy | The success of the solution is contingent upon public confidence in the system and a perception of its legitimacy. A lack of trust could hinder adoption and effectiveness. | A user survey-based metric, such as a Net Promoter Score (NPS), to gauge public trust and satisfaction with the system. |
| Innovation and Continuous Improvement | The implementer must have a demonstrated commitment to innovation, with a clear incentive structure for ongoing system enhancements and continuous improvement. | Evidence of innovativeness and a history of successful technological improvements in past projects. |
The involvement of private players is significantly beneficial in three criteria, namely a) Scalability & Technological Expertise, b) Cost Efficiency and c) Innovation and Continuous Improvement. Unlike government agencies, private firms possess specialised knowledge and are incentivised for continuous improvement and efficiency.
Despite the benefits of private involvement, the current legal framework and public trust considerations necessitate continued government oversight which provides the necessary balance. While a private entity can provide the technology, the government’s legitimacy is essential for the legal integrity and public confidence in the FIR registration process.
The following image shows the comparison of the three approaches across the criteria for the composition of the implementation stakeholder:
Comparison of the Three Approaches
Based on the above image, a public-private partnership has clear benefits over a pure public endeavour in incorporating technology into the FIR registration process. The complete privatisation of the FIR registration process is not a viable option under the current legal and institutional framework. The police’s role in supervising and authenticating FIRs is a core state responsibility defined by existing laws, which prohibits the transfer of this duty completely to private actors.
6. Process
The objective has been defined with clearly measurable targets as follows:
To deploy a scalable, multilingual, AI-assisted FIR registration system across two districts (one urban and one rural) within 17 months, integrating e-KYC, voice-to-text, Aadhaar e-signatures, case dashboards, and LLM-based FIR drafting with human review. The system should achieve 90 per cent FIR accuracy, a 70 per cent reduction in officer documentation time, 85 per cent satisfaction among victims and officers, and 95 per cent SLA compliance in registration.
The solution will be deployed in four phases. An overview is shown below:
Overview of the Solution
The detailed breakdown of each phase, including its time-frame, core activities, desired outcomes, and success criteria, is provided below. Advancement to a new phase is contingent upon meeting the success criteria for all stages within the preceding phase.
| Stage | Core Activities | Key Outcomes | Success Criteria |
|---|---|---|---|
| Phase 1 | |||
| Stakeholder Alignment (4 months) | Constitute three 3-tier steering committee
Finalise scope, KPIs, escalation matrix Sign MoUs |
|
Budget released (expected to be ~5 crores for 2 districts) Roles accepted by all parties |
| Research and On-Ground Process Mapping (1.5 months) |
|
Functional requirements spec finalisation | Steering committee sign‑off on To‑be SOP |
| Phase 2 | |||
| Solution Design & Vendor On‑boarding (2.5 months) |
|
|
Signed Contracts with vendors CERT‑In approval on security specs |
Product Validation & Benchmarking (2 months) |
|
Validation report (≥90 % accuracy on dev set) | Independent QA sign off |
| Phase 3 | |||
| Integration (2 months) |
|
|
All high-severity defects closed Security clearance post integration |
Training Sessions (2 months) |
|
|
Training sessions FAQ Portal deployment |
| Phase 4 | |||
Pilot Roll‑out (3 months) |
|
|
KPIs: 90% accuracy, 25 % faster filing by Week 8 from rollout Infosec clearance |
Scaling Proposal (2 months) |
|
|
Govt. green‑light scaling |
In Phase 4, the results from the pilot will be audited, findings analysed and if there is a successful outcome, a proposal for a statewide roll-out will be prepared.
6.1. Challenges
The implementation of the pilot phase is anticipated to face several challenges, which are outlined below along with their corresponding mitigation strategies:
| Challenge | Description | Mitigation Strategy |
|---|---|---|
| User Unfamiliarity | A segment of the population may find the technology driven process daunting despite increasing digital literacy. | It is crucial to ensure that human handlers are adequately trained and sensitised to the issues that informants may face. Introducing chatbots can also familiarise users with the technology and the FIR filing process. |
| Resistance from State Governments | State governments may express concerns regarding implementation costs, potential job displacement, or a general reluctance to adopt technology prematurely. | Evidence from successful international and domestic use cases should be provided to the state Governments. It is critical for the government to be aligned that the technology shall enable police officers to prioritise core law enforcement duties by offloading administrative tasks. |
| Privacy and Data Protection | The integration of new technology into traditional, human-driven processes will inevitably raise questions about data privacy and security. | A robust technical architecture with additional information security checks will enhance privacy and data protection. This ensures that the new system is an improvement over existing safeguards. |
| Connectivity Issues | The new system’s efficacy relies heavily on stable and reliable internet connectivity across all operational areas. | Connectivity is already a critical requirement for modern policing. If necessary, additional investment shall be needed to develop a seamless network infrastructure across all police stations. |
| Spurious FIRs | There is a concern that increased automation could lead to a surge in the number of FIRs, including a higher volume of spurious or fraudulent complaints. | For non-cognizable offences and non-investigative cases, such as lost items, this issue should not be a major concern. For cognizable offences, since the ultimate authority for FIR registration resides with a police officer, this issue can be mitigated. |
7. Conclusion
This policy brief outlines a strategic framework for modernising the First Information Report (FIR) filing process in India through the integration of AI technology. After evaluating various models of technological integration, a hybrid approach that leverages a public-private partnership (PPP), is recommended. This model is designed to automate key aspects of FIR registration while retaining essential human oversight and judgement.
The successful implementation of this hybrid model requires a phased approach. A pilot project is proposed in one rural and one urban district within a single state, to test the efficacy of the recommended approach and gather insights from diverse operational environments.
The implementation process must be guided by a clear, measurable, and time-bound plan. Each stage must have predefined metrics to assess performance and progress. If the pilot project successfully meets its intended outcomes, a comprehensive plan can subsequently be developed to scale the initiative across the entire state.
The adoption of AI and other advanced technological solutions is crucial to the modernisation of India’s police force. This integration drives significant improvements in the delivery of law enforcement services for citizens, while also enhancing its operational efficiency, improving outcomes and state capacity, and easing the administrative burden on an already pressurised police force. A carefully designed and implemented model, embracing technological evolution anchored in human supervision, will be a vital step toward creating a more efficient, transparent, and responsive law enforcement system built to capably handle the challenges of today and of the future.
Footnotes
Comptroller and Auditor General of India. 2022. “Report No. 04 of 2022.” cag.gov.in.↩︎
Lalita Kumari v. Government of Uttar Pradesh, Supreme Court of India, November 12, 2013, (2013) 2 SCC 1.↩︎
Panigrahi, Debasish. “Maharashtra Police Stations Don’t File FIRs in 50% of Cases, Shows Internal Survey.” Hindustan Times. December 9, 2017.↩︎
Gera, Ishaan. “Indian Police Forces Struggle with Limited Computers despite Rising Digital Crime: Moneycontrol Analysis.” Moneycontrol. August 29, 2024.↩︎
Ministry of Home Affairs, Government of India. “Answer to Unstarred Question No. 3266 in the Rajya Sabha on March 24, 2021.” New Delhi: Ministry of Home Affairs. 2021.↩︎
India Justice Report 2022: National Factsheet, Tata Trusts and India Justice Report (New Delhi: India Justice Report, 2022)↩︎
James, Nathan. “The Community Oriented Policing Services (COPS) Program: History, Funding, and Potential Issues for Reauthorization”. Congressional Research Service. April 24, 2025.↩︎
Lalita Kumari v. Government of Uttar Pradesh & Others, Supreme Court of India, November 1, 2023, (2014) 2 SCC 1.↩︎
Comptroller and Auditor General of India. Information System Audit of Implementation of “Crime and Criminal Tracking Network Systems” (CCTNS). New Delhi: Indian Audit and Accounts Department, n.d.↩︎
Comptroller and Auditor General of India. “Report of the Comptroller and Auditor General of India on Social, Economic and General Sectors. Government of Assam (Report No. 04 of 2022).” Comptroller and Auditor General of India.↩︎
Comptroller and Auditor General of India. “Report of the Comptroller and Auditor General of India on Social, Economic and General Sectors. Government of Assam (Report No. 04 of 2022).” Comptroller and Auditor General of India.↩︎
Government of India. Bharatiya Nagarik Suraksha Sanhita, 2023. Section 173. New Delhi: Ministry of Law and Justice, 2023.↩︎
“Common Service Centre Program”, Digital governance.” Vikaspedia. May 9, 2023.↩︎
Barus, Becca. “ChatComm Is A Private Center Handling Public Response.” The Journal of Emergency Dispatch. January 14, 2020.↩︎
“Common Service Centre Program”, Digital governance.” Vikaspedia. May 9, 2023.↩︎