According to The Times of India, Amazon's global trust and safety chief Rohan Oommen has identified a critical challenge facing India's e-commerce sector: protecting millions of first-time online shoppers from AI-powered scams. Speaking to TOI, Oommen emphasized that consumer awareness has become as critical as technology in tackling digital fraud. The rapid influx of new digital commerce users from tier-2 and tier-3 cities has made India distinct from more mature markets, where e-commerce adoption has largely plateaued.
The Indian government has launched comprehensive deepfake detection projects to address AI-generated content challenges. Three projects were selected under the second call for proposals within the Safe & Trusted AI program, including Saakshya developed by IIT Jodhpur and IIT Madras - a multi-agent, retrieval-augmented framework for deepfake detection and governance. AI Vishleshak, led by IIT Mandi and the Directorate of Forensic Services in Himachal Pradesh, focuses on improving detection of audiovisual deepfakes and handwritten signature forgeries. IIT Kharagpur's project targets synthetic speech and voice cloning attacks in real time. These initiatives complement the fourth edition of Mission GraHAQ, launched on July 31, which pairs consumer education with automated threat detection.
As reported by The Times of India, generative AI has made fraud more sophisticated by enabling fake identities, forged documents and synthetic reviews. However, Oommen noted that generative AI has also strengthened Amazon's ability to detect abuse at scale. The company is increasingly deploying AI systems, including agentic AI that can carry out continuous audits, while relying on years of behavioural data to distinguish genuine sellers from bad actors. The government's regulatory response includes amendments to the Information Technology (IT) Rules that introduced labeling, traceable metadata, and platform checks for AI-generated content.
Amazon has significantly upgraded its fraud detection capabilities with the launch of SENTRIX AI infrastructure for 2025. According to Unbox Daily HQ, this backend system actively identifies malicious phishing websites within 24 hours and flags fraudulent phone numbers on the same day. The company continues to deploy Claude AI to Indian sellers at zero cost, while maintaining its focus on behavioural patterns rather than documents or images as the most reliable method for distinguishing genuine sellers from fraudulent actors. The government has also conducted 6,650 cybersecurity awareness workshops for more than 1.13 million participants, including students, teachers, law enforcement personnel, government employees, and members of the public.
Visa has announced its acquisition of BioCatch, a leading fraud prevention company that uses human behavioral patterns to detect financial crime. BioCatch's technology collects over 3,000 anonymized data points including keystroke and mouse activity, touch screen behavior, AI agent usage, and device characteristics to assess user intent in real-time. The company protects more than 760 million users across 1.8 billion devices through its solutions deployed by more than 350 financial institutions. This acquisition represents Visa's strategic move to strengthen its fraud detection capabilities as AI-powered financial crimes continue to evolve.
