AI Debt-Collection Platform Leaks 6.6 Million Accounts and 315 K Call Recordings
Two critical bugs in CreditNirvana's AI-driven debt collection system exposed 6.6 million borrower accounts and 315,000 call recordings, prompting urgent security fixes.
S
··1 min readSciTech Editorial Team
Newsroom

An AI‑driven debt‑collection platform, CreditNirvana, exposed the personal data of 6.6 million borrowers and roughly 315,000 call recordings after two serious vulnerabilities were discovered. The breach, revealed by independent researcher Eaton Zveare, underscores how financial‑service firms that rely on autonomous decision‑making tools can become high‑value targets for data theft.
The first flaw, classified by India’s Computer Emergency Response Team (CERT‑In) as “high” severity, affected the platform’s call‑center module. It allowed an attacker to retrieve hundreds of thousands of recorded conversations and the AI‑generated recommendations that followed each call. Moreover, the bug could have been leveraged to commandeer the platform’s autonomous agents, redirecting them to launch bespoke collection campaigns.
A second, “critical” severity vulnerability struck CreditNirvana’s core engine. It leaked the records of three major corporate clients—prominent Indian financial‑services firms—as well as the accounts of millions of individual borrowers. Exposed data included credit reports, identity documents, case‑management notes, court filings, and a trove of emails, text messages, and call transcripts related to unpaid debts. CERT‑In warned that exploitation could let a malicious actor enumerate users, assume arbitrary sessions—including administrative ones—and modify sensitive information with full privileges.
The defects were patched in June 2024 after Zveare coordinated the disclosure with CERT‑In, and the fixes were publicly announced in October. While the rapid remediation earned praise, Zveare cautioned that the bugs may have persisted for years, leaving a massive amount of personal finance data vulnerable. Both CreditNirvana and its parent fintech, Perfios, declined to comment on the incident.
The episode arrives at a moment when AI is being embedded into ever‑more automated financial workflows. It serves as a stark reminder that the convenience of autonomous debt‑collection must be balanced with rigorous security audits and robust regulatory oversight. Industry observers predict that tighter standards for AI‑driven fintech services will emerge, aiming to prevent similar exposures that could erode consumer trust and invite costly legal repercussions.



