AI Regulation Should Be Rational, Not Retaliatory

• The Trump administration is accused of using AI regulation as a tool for retaliation against Anthropic, rather than implementing rational national security policy.
• EFF and other groups filed an amicus brief arguing that government sanctions on the company are unconstitutional and violate First Amendment principles.

Retaliation Against Anthropic

• Government officials reportedly targeted Anthropic after the company refused to allow the Pentagon to use its AI models for autonomous weapons systems and domestic surveillance.
• The administration designated Anthropic a "supply chain risk," barring contractors from working with it; a court-issued preliminary injunction currently blocks these sanctions.
• New executive orders imposed export controls specifically on Anthropic’s "Mythos" and "Fable" models, forcing the company to shut them down.

Constitutional and Security Implications

• The EFF argues that export controls on software tools like LLMs mirror 1990s battles over encryption, reaffirming the legal principle that "code is speech."
• The administration’s move is seen as arbitrary, as other LLMs with similar cybersecurity capabilities remain subject to only voluntary, light-touch testing frameworks.
• Experts warn that these restrictions actually hinder security by preventing researchers and teams from using advanced models to identify and patch vulnerabilities before adversaries do.