AI Debt-Collection Service CreditNirvana’s Security Gaps Threaten Data of 6.6 Million Users
Research uncovered two serious flaws in CreditNirvana, the AI‑powered debt‑collection platform, that might have allowed attackers to obtain personal information belonging to approximately 6.6 million people. The same weaknesses also jeopardized close to 315,000 stored call recordings, as reported initially by Wired.
The service collects extensive consumer data to automate contact regarding delinquent accounts. Among the exposed records are email addresses, text messages, and complete credit reports—files that disclose balances, loan histories and payment patterns. Exploiting the bugs could have let a hacker pull this data without any alerts or authentication hurdles.
Researchers reviewing the software found that the API did not apply adequate authorization on two separate endpoints. One endpoint unintentionally offered a bulk‑download capability for account metadata, and the other permitted access to saved call recordings via a predictable URL pattern. Both defects arose from weak input validation and missing role‑based access controls—issues frequently seen in fast‑tracked fintech deployments.
The possibility of a breach alarms both consumers and regulators. In the United States, the Fair Credit Reporting Act safeguards credit‑reporting information, and illegal exposure can attract hefty fines. Additionally, the massive volume of compromised records could enable identity‑theft operations, phishing campaigns, and other fraudulent activity. Advocacy organizations have called on CreditNirvana to inform the impacted individuals and work with officials to determine whether the data has been misused.
CreditNirvana replied that it has remedied the discovered flaws and is undertaking a thorough security audit. The firm also said it is partnering with external cybersecurity specialists to confirm that no breach took place prior to the patch. As the probe proceeds, analysts note that the case highlights the need for stringent security testing of AI‑driven financial applications, particularly those that process sensitive personal data.
Comments (0)
Be the first to comment.
Join the discussion