Anthropic, the company behind the AI chatbot Claude, has warned users about a data theft campaign involving malware designed to steal login information and drain account credits. The malicious software, including variants like Vidar, Lumma, StealC, RedLine, and Atomic Stealer, is typically found on Windows and Mac computers. These programs are known as "infostealers," meaning they are designed to capture sensitive data such as passwords and financial details. There is currently no evidence that mobile devices are affected by this particular campaign. In an email to users, Anthropic explained that the malware is being used to access Claude accounts and exhaust their usage credits. This could lead to unexpected charges for users who are not aware of the unauthorized activity. Those affected have been logged out of their accounts, and their payment information has been removed. The company has begun issuing refunds for any unauthorized usage fees, but it clarified that it is not responsible for charges resulting from users downloading pirated software or engaging in illegal activities. A Reddit user shared a screenshot of the email sent by Anthropic, which suggested that the malware may have come from pirated software, such as a pirated video game. The company emphasized that the malware is not connected to Claude or any activities performed on the platform. Users are advised to remove any suspicious or pirated software from their systems and run a full system scan to detect and remove any information-stealing malware. Anthropic has taken steps to protect users' financial data and is refunding unauthorized usage fees. Affected users are encouraged to log back into their accounts and re-enter their payment information. If users notice any unusual fees that have not been refunded, they are advised to contact Claude’s support team for further assistance. The company continues to monitor the situation and recommends that users remain vigilant about the software they download and install.