Ask HN: Anthropic banned me from using Claude Code and I don't know what to do

The Human Cost of AI Safety Automation

A developer’s experience on Hacker News highlights a growing friction point in the AI industry: automated enforcement systems that lack human judgment. After using Claude Code at work for months, this user attempted to use the tool on personal projects, only to be banned within an hour—likely triggered by a VPN connection. Despite receiving a refund, a second attempt a month later (this time without a VPN) also resulted in a ban, presumably because the same credit card was reused. The user now faces a permanent account suspension with no clear path to resolution.

Why This Matters

Anthropic, like many AI companies, employs aggressive anti-abuse systems to prevent fraud, credential stuffing, and policy violations. These systems are necessary—AI tools are expensive to run and vulnerable to misuse. However, the case illustrates a classic problem: automated bans often lack nuance, catching legitimate users in the crossfire. The developer’s VPN usage, common for privacy-conscious professionals, flagged the account. The subsequent ban on re-registration with the same payment method shows that identity-based blacklisting can create a permanent shadowban for users who made an honest mistake.

This isn’t just an inconvenience—it’s a reputational risk for Anthropic. Developers are the primary audience for Claude Code, and alienating them through opaque enforcement erodes trust. When a user who has been a paying customer for months cannot even access customer support to appeal, the message is clear: the system prioritizes risk mitigation over user experience.

Implications for AI Practitioners

For developers and AI practitioners, this incident serves as a cautionary tale. First, VPN usage with AI services is risky—many platforms treat it as a red flag for fraud. Second, once banned, recovery is difficult; shared payment methods, IP addresses, or even browser fingerprints can link accounts. Third, the lack of a human appeals process means users are at the mercy of automated systems that may never reconsider.

This also raises questions about how AI companies handle false positives. As AI tools become more embedded in professional workflows, a permanent ban can disrupt projects, break CI/CD pipelines, and damage productivity. The industry needs better escalation paths—perhaps a “trusted user” status for long-term customers, or a simple verification process to lift bans.

Key Takeaways

• Automated enforcement systems can permanently lock out legitimate users due to false positives from VPNs or shared payment methods, with no clear appeal process.
• Developers should avoid using VPNs with AI coding tools unless explicitly permitted, as many platforms treat them as fraud indicators.
• AI companies need better human-in-the-loop support for account bans, especially for paying customers with established usage histories.
• Practitioners should diversify their tooling to avoid single-vendor lock-in, as a sudden ban can halt critical workflows with no recourse.