The ability to approach the police and seek justice remains difficult for many citizens. Administrative procedures, documentation requirements, language barriers, and varying levels of literacy continue to shape how individuals engage with the criminal justice system.
As governments increasingly explore the use of Artificial Intelligence (AI), an important policy question emerges: can technology help make justice systems more accessible to citizens?
The experience of DHARMA, an AI-enabled policing initiative implemented by Eluru Police in Andhra Pradesh, provides a useful lens through which to examine this question. Its significance lies not only in the adoption of new technology, but in the broader challenge it seeks to address: improving access to policing services in a diverse and unequal society.
Accessibility as an Institutional Challenge
The first point of engagement with the police is the reporting of a grievance. This act requires individuals to navigate formal procedures, often at moments of distress, uncertainty, or vulnerability. For many citizens, the difficulty lies not only in approaching the police, but in knowing what information is relevant, how it should be communicated, and what consequences may follow once a complaint is recorded.
These challenges extend beyond the citizen's experience. Frontline personnel often spend considerable time translating narratives into formats that satisfy procedural and legal requirements. As information moves from citizens to records, and from records to investigation, the quality and clarity of that information become critical to institutional functioning.
Improving accessibility, therefore, is not solely a matter of citizen convenience; it is also an operational concern that influences the efficiency and effectiveness of policing services.
DHARMA: An Experiment in Accessible Policing
DHARMA was designed to address a practical problem within policing: how to improve the recording of citizen complaints while reducing administrative burdens on frontline personnel. By enabling citizens to communicate in natural language while supporting the conversion of those narratives into structured administrative records, the platform strengthens both access at the point of entry and the quality of information entering the policing system.
DHARMA's deployment is accompanied by safeguards designed to ensure that AI supports, rather than substitutes for, human judgement. Police personnel are trained to use the platform, and all AI-generated outputs, including complaint summaries, FIR drafts, and investigation recommendations, are reviewed and approved by the concerned police officer. The system also incorporates measures relating to data security, auditability, and evidence management, ensuring that technological assistance remains embedded within established legal and administrative processes. These safeguards have been important in enabling the adoption of AI while preserving institutional accountability.
The platform has processed more than 2,000 complaints and supported the training of over 150 police personnel in the last one year. FIR drafting time has reduced from approximately 30 minutes to 12–15 minutes per case, indicating the potential of AI-assisted workflows to reduce administrative effort at the frontline. This has resulted in improvements in documentation quality, case disposal, and conviction outcomes.
Efficiency and quality gains at the stage of complaint recording have larger institutional implications. Eluru district has 2,000police personnel serving an estimated 2.1 million people. In such a setting, where administrative capacity must serve a large population, DHARMA then becomes a pathway to reduce the distance between citizens and the state at one of the most consequential points of public engagement: the search for justice.
Beyond Eluru: Designing Accessible Justice Systems
The significance of DHARMA extends beyond a single district or technology platform. Voice-based interactions, multilingual interfaces, assisted documentation, and citizen-centred workflows represent design choices that seek to align institutional processes more closely with how people communicate and seek help. As governments continue to invest in digital public infrastructure, such considerations become increasingly relevant across a range of citizen-facing services, including grievance redressal, welfare delivery, legal aid, and municipal administration.
Scaling such approaches, however, requires more than technological deployment. Training shapes how frontline personnel engage with new systems, while integration determines whether information can move seamlessly across existing processes. Where these elements align, technology can reinforce institutional capacity. Where they do not, even promising innovations risk adding complexity instead of reducing it.
Questions of reliability and accountability become equally important as adoption expands. Systems that influence documentation and decision-making must produce outputs that can withstand scrutiny, support established procedures, and retain public confidence. Scaling is therefore not simply a question of expanding access to a tool; it is a question of embedding new capabilities within institutional practice.
Bringing Citizens Closer to the State
Public institutions are ultimately experienced through moments of interaction. The ease with which a grievance can be communicated, recorded, and acted upon shapes how citizens encounter the state in their everyday lives. If initiatives such as DHARMA succeed in reducing this distance, their significance will extend beyond technology. They will demonstrate how improvements in accessibility can strengthen the relationship between citizens and public institutions, particularly at moments when people seek protection, assistance, and justice.