Citing a 2025 NITI Aayog assessment, the report estimated that poor-quality data causes leakage of 4-7% of India’s annual welfare spending.

NEW DELHI: Even as India processes over 20 billion UPI transactions every month and has enrolled more than 1.4 billion people under Aadhaar, fragmented government databases continue to cause financial losses, exclude eligible beneficiaries from welfare schemes and threaten the country’s AI ambitions, according to a report released on Thursday.The Tony Blair Institute for Global Change (TBI) report, From Digital Scale to Data Power: A Data Operating Model for India’s States, said India’s next digital challenge is not building more infrastructure but making existing government data accurate, accessible and usable across departments.Citing a 2025 NITI Aayog assessment, the report estimated that poor-quality data causes leakage of 4-7 per cent of India’s annual welfare spending.It highlighted how past data-cleaning exercises helped remove 1.71 crore ineligible PM-KISAN beneficiaries, saving an estimated Rs 9,000 crore. Similarly, eliminating 3.5 crore bogus LPG connections saved Rs 21,000 crore over two years, while removing 1.6 crore fake ration cards is saving approximately Rs 10,000 crore annually.

Despite these gains, government databases covering health, education, land records, taxation and welfare largely operate independently, limiting their usefulness for policy decisions.The report warned that deploying AI on fragmented or unreliable databases could amplify existing errors rather than improve public services.“Is the data good enough to trust an AI system, or a human official, to act on it? Right now, in most states, the honest answer is not yet,” said Vivek Agarwal, TBI country director and report co-author.Drawing on initiatives in Karnataka, Odisha and Rajasthan, the report proposed seven reforms, including common data standards, independent quality checks, stronger accountability and legal safeguards.It also recommended a State Data Balance Sheet to help governments track data assets and risks, arguing that reliable information must precede large-scale AI deployment in governance.