Low-Cost Open-Source AI Models Are Powering India's Rapidly Escalating Deepfake Fraud Wave, Overwhelming KYC Defenses

India's worsening deepfake fraud epidemic is being driven not by frontier AI but by cheap wrappers built on open-source models, shared via Telegram and dark-web channels and runnable on consumer gaming hardware at minimal cost, according to Business Standard reporting June 17. Total Indian bank fraud losses rose 46.4% year-over-year to ₹48,021 crore (~$5.7B) in FY26 despite fewer total cases—pointing to AI-enabled synthetic-identity attacks on live video KYC checks pushing up the per-incident loss figure. A compounding risk: AI deepfake detection tools produce a 61.2% false-positive rate against non-native English speakers versus 5.1% for native speakers, exposing legitimate customers to wrongful fraud accusations and creating legal liability for deployers.