After Anthropic Tightens Claude Access, Users Test Gemini 4 as Alternative

On October 8, 2026, Anthropic published its 2026 update to the Claude usage policy, clarifying the Acceptable Use Policy (AUP) and the rules governing supported regions. By November 12, the measures behind what observers have described as Claude's strictest account-enforcement regime took effect, introducing penetration-style review criteria for regional access with the stated aim of shutting down shell companies and proxy relays used to bypass location restrictions.

The rules apply across the full Claude product line: Claude.ai, Claude Code, and Cowork, the API and developer platforms, and Claude model services delivered through cloud providers and authorized resellers, all of which now fall under one unified set of regional-access and usage constraints. Access is restricted for three categories of users. First, anyone physically located in an unsupported region is barred from using the service even if employed by an organization in a compliant region. Second, an operating entity registered or headquartered in a restricted region loses access regardless of where its staff are located. Third, companies that are directly or indirectly held or effectively controlled by entities based in restricted regions are denied access, no matter where they are registered or where the end user sits.

Beyond regional rules, the policy adds constraints for risky use cases. It prohibits running networks of fake accounts, building fake news sites, and conducting commercial or opinion-manipulation campaigns with Claude, and it bars the model from 11 high-risk decision-making scenarios covering areas such as healthcare, finance, law, employment, credit, and insurance. The mechanism is now in force, and no further changes have been disclosed.