Indian regulator uses AI to scan social media for investment fraud

The Securities and Exchange Board of India (SEBI) is now utilizing an artificial intelligence system to scan public social media for fraudulent investment claims, unregistered financial advice, and the impersonation of regulated entities. SEBI disclosed this capability in its Annual Report 2025-26, detailing that its AI system, Project SUDARSAN, became operational in November 2025 and has already identified over 20,000 instances of potentially illicit content.

Project SUDARSAN, which stands for Surveillance of Unauthorized Digital Activity via Real-time Scanner for Anti-fraud, employs multimodal AI to analyze speech, visuals, regional languages, and context from publicly available content across major social media platforms. The system assigns risk scores by combining this analysis with behavioral and regulatory parameters, flagging potential violations for further scrutiny. It is designed to detect promises of guaranteed returns, fake certifications, and instances where investment advice is given without proper registration.

This initiative arrives amidst growing concerns about the influence of social media "finfluencers" on investment decisions. SEBI's Investor Survey 2025 indicated that 62% of investors are influenced by finfluencers, who are largely perceived as credible. YouTube, Instagram, and Facebook were identified as the most frequently used platforms for seeking securities market information.

In parallel, SEBI has deployed another AI tool, R(AI)DAR, to review advertisements and investor education materials from asset management companies, flagging potential violations such as missing disclosures. These technological advancements represent a strategic shift for SEBI, moving from manual, complaint-driven supervision towards proactive, technology-based monitoring to detect harmful content earlier.

However, the annual report lacks details on SUDARSAN's risk scoring methodology, the thresholds for regulatory scrutiny, its handling of false positives, and the safeguards for data collection and retention. While SEBI states the system only scans public content related to securities fraud, it does not specify the platforms covered or data retention periods. The absence of transparency regarding these aspects, along with safeguards for automated risk scoring and user recourse, raises questions about oversight and the limits of automated surveillance, echoing past controversies surrounding government proposals for online monitoring.