Anthropic Banned 11.4 Million Claude Accounts in Six Months with Low Appeal Success Rate

A transparency report from Anthropic revealed that the company banned 11.4 million Claude accounts in the first half of the year. Out of nearly 398,000 submitted appeals, only 42,000 accounts were successfully reinstated, yielding a low success rate of approximately 10.5%. As developers and organizations rely heavily on Claude for daily coding and complex workflows, aggressive automated bans risk wiping out critical productivity assets overnight. This highlights the growing friction between strict platform trust and safety enforcement and user dependence on cloud-hosted AI infrastructure. Anthropic enforced account bans using in-house machine learning classifiers, hash matching, and automated regional restriction checks, while submitting over 15,000 extreme content violation reports to safety organizations like NCMEC. Affected developers complained that the appeal process relies on automated rejection templates with no clear explanation of violated terms or avenues for human review.