Home / Technology / AI Chat Hijacked: Malware Steals Claude Sessions
AI Chat Hijacked: Malware Steals Claude Sessions
2 Sep
Summary
- New malware targets AI platform credentials, stealing Claude sessions.
- Infostealers like Lumma and RedLine compromise Windows and Mac PCs.
- Users urged to clean devices; Anthropic refunds unauthorized charges.
A new wave of cyber threats is targeting AI platform users, specifically those utilizing Claude. Malicious actors are employing infostealer malware to compromise user sessions, with the primary goal of stealing usage credits. This campaign has been detected on both Windows and Mac operating systems, though mobile devices are not currently believed to be affected.
Several known infostealers, including Vidar, Lumma, StealC, and RedLine on Windows, and Atomic Stealer on a small number of Macs, are being utilized in this attack. These threats quietly infiltrate accounts, consuming user allowances and potentially leading to unexpected charges. Anthropic has acknowledged the issue, signed out affected users, and removed payment details to prevent further abuse. The company is also working to refund unauthorized usage charges.
Anthropic has clarified that this malware is not related to Claude itself but rather exploits user habits, such as downloading cracked software or pirated material. This serves as a critical reminder for users to exercise caution with all software downloads. Comprehensive system scans and removal of suspicious or cracked software are essential steps to safeguard devices and personal data from these persistent threats.
