INSIGHT · GLOBAL
Credential theft of AI service subscriptions enables unauthorized model access at scale
Attackers are compromising login credentials for paid AI service subscriptions (ChatGPT Plus, Claude Pro, etc.) and reselling access. This allows free/low-cost model usage and creates account-takeover risk for financial services employees.
WHY IT MATTERS
Banks deploying enterprise LLM access must enforce MFA, monitor anomalous API usage, and audit model interactions for sensitive data leakage. Compromised subscriptions can bypass approval workflows and expose proprietary data.
Source: NewsBytes · 2026-09-27