AI Fraud Detection Tools Unveiled at Ayushman Bharat Hackathon to Curb ₹630 Crore in Fake Claims

Loading article...

AI Fraud Detection Tools Unveiled at Ayushman Bharat Hackathon to Curb ₹630 Crore in Fake Claims | AirNews

The National Health Authority (NHA) has flagged the rising use of AI-generated fake medical records and forged insurance claims as a critical threat to the Ayushman Bharat Pradhan Mantri Jan Arogya Yojana (AB-PMJAY), as officials unveiled new AI-driven fraud detection tools developed during a national hackathon held in Bengaluru on Saturday. The event, organized with the IndiaAI Mission and the Indian Institute of Science (IISc), targeted solutions for identifying document forgery, deepfakes, and medically implausible claims.

NHA CEO Sunil Kumar Barnwal stated that existing anti-fraud systems have already prevented fraudulent claims worth ₹630 crore, with an additional ₹200 crore recovered or penalized from hospitals. Joint Secretary Jyoti Yadav highlighted that AI-generated clinical notes, discharge summaries, and manipulated diagnostic reports are now being submitted, often involving altered images, removed watermarks, or copy-pasted content. The hackathon sought models to verify document authenticity, classify medical records, and ensure submitted reports align with claimed treatments.

Several innovative solutions emerged, including a fraud-detection model based on mathematical analysis rather than conventional AI. Officials cited cases of minor ailments reported as major surgeries, reused ICU patient photos, and chemotherapy billed at medically impossible intervals. Machine-learning tools are now being trained to validate lab parameters against treatment claims, such as checking creatinine levels for dialysis eligibility. Data analytics also revealed cross-State patterns where clusters of beneficiaries sought identical treatments shortly after card issuance.

The NHA emphasized the need for balanced automation, warning that overly rigid systems could impact legitimate claims due to medical exceptions. Fraud patterns include inflated billing, fabricated records, and exploitation of higher reimbursement rates for ICU admissions. The authority is evaluating the hackathon's winning models for integration into the national claims processing framework.

The NHA will finalize the deployment framework for the selected AI tools by June 2026, with pilot implementation expected in high-fraud districts by July.