Hackers are stealing Claude tokens from subscribers

Anthropic has issued a warning to users following reports of unauthorized activity on Claude accounts. The issue came to light after a subscriber noticed significant token consumption despite not actively using the service. Investigations suggest that malicious actors are targeting subscriber accounts to hijack their API tokens, potentially for unauthorized access to the AI model's capabilities. This security incident highlights the growing risks associated with generative AI platforms, where stolen credentials can be leveraged to bypass usage limits or access sensitive data. Anthropic has advised users to remain vigilant, monitor their account activity for irregularities, and implement robust security measures, such as enabling multi-factor authentication, to protect their accounts from compromise. The company is currently working to address the vulnerability and mitigate the impact on affected users, emphasizing the importance of account security in the rapidly evolving landscape of AI-powered services.