AI Account Hacked: Unauthorized Charges in Euros Linked to Claude App

A Bay Area man, Jason Weitzman, reported unauthorized charges totaling around $315 on his Anthropic Claude account, which he claims were made without his consent. The charges were flagged by Chase Bank, prompting Weitzman to investigate. He discovered a problematic add-on, or 'skill', connected to Claude that used his stored payment information to make purchases. Anthropic stated that the issue stemmed from a compromise of Weitzman's device, not the Claude app itself. After multiple attempts to resolve the issue with Anthropic, Weitzman sought help from 7 On Your Side, which facilitated a resolution. Anthropic later refunded Weitzman and restored his access to Claude. The incident highlights potential vulnerabilities in AI applications and user device security.

Key Points: • Unauthorized charges of approximately $315 were made on a user's Anthropic Claude account. • The issue was linked to a compromised device rather than a flaw in the Claude app itself. • User support from Anthropic failed until media intervention prompted resolution and a refund.