Data and justice: On courts in India and AI tools

The introduction of OCOD and Su-Sahayak is a significant step in the ongoing digitisation of the Indian judiciary, driven by the e-Courts Project. OCOD aims to create a unified digital trail for disputes, addressing the long-standing issue of fragmented records across district and subordinate courts. This standardized data can enhance judicial administration by identifying procedural bottlenecks and improving data-based decision-making. However, the Supreme Court must navigate the complex landscape of interoperability between different courts and the integration of legacy records. The success of OCOD hinges on overcoming these administrative hurdles and ensuring a seamless flow of information across the judicial hierarchy.

While the tools are designed to improve "access to justice", a core constitutional principle under Article 39A, they risk exacerbating the digital divide. OCOD may require technological upgrades (scanners, software, cloud storage) that are financially burdensome for independent practitioners at the district and taluka levels, potentially creating a disparity between well-resourced corporate firms and grassroots lawyers. Furthermore, the reliance on digital platforms could introduce unregulated "digital middlemen" for litigants unable to navigate e-filing portals. The text-based nature of 'Su-Sahayak' may also exclude individuals with low digital literacy, highlighting the need for voice-assisted features like those in the government's Jan Sahayak initiative. The state must ensure these reforms do not inadvertently erect new barriers to justice.

The deployment of AI tools like 'Su-Sahayak', following the earlier introduction of SUVAS (for translating judgments) and SUPACE (for processing facts and legal precedents), reflects the judiciary's cautious adoption of technology. Currently, AI is primarily used for administrative assistance rather than substantive legal reasoning. This distinction is crucial, as algorithms can perpetuate existing societal biases. The article notes the critical need to ensure AI models are not biased against marginalized communities, who historically face disproportionate arrest and bail denial rates. A centralized digital fingerprint for each case under OCOD also raises concerns about data privacy and the potential misuse of sensitive information, necessitating robust data protection frameworks aligned with the Digital Personal Data Protection Act, 2023.