Two weaknesses in an AI-driven debt collection platform exposed 6.6 million accounts and about 315,000 call recordings. Accessible information reportedly included credit reports, identification documents, legal records, messages, transcripts and details about unpaid accounts. One issue could also have enabled arbitrary user sessions, including administrative access. The weaknesses were fixed in June and later disclosed publicly.
Why It Matters
Borrowers faced potential exposure of highly sensitive financial and personal histories. Three prominent Indian financial-services customers were reportedly affected through the platform. For leaders: AI services can concentrate information from many systems into one high-impact exposure. For defenders: assess vendor access, administrative impersonation controls and the retention of recordings and documents. The larger lesson is that automation increases the need for strict data minimization and access boundaries.
What Defenders Should Do
Review AI service providers for excessive data retention, cross-customer exposure and controls that prevent user or administrator impersonation.
