The concept of “expressive governance” — where government agencies use administrative tools to penalize companies for their speech on artificial intelligence matters — is facing scrutiny in a federal court case. The legal challenge comes after a leading U.S. AI company was designated a “supply-chain risk to national security” and barred from federal defense contracts due to its stance against the use of its AI models in autonomous weapons and mass surveillance. Critics argue this tactic suppresses protected speech under the First Amendment by cloaking retaliatory measures in routine regulatory actions.
What Happened
In the ongoing case Anthropic v. US Department of War before a San Francisco federal court, the government labeled the AI firm as a security risk, citing concerns over its refusal to allow its technology’s use in autonomous weapons and surveillance on U.S. citizens. Following this, the President directed all federal agencies to cease using the company’s AI technology across the government’s defense ecosystem. This government action did not arise from a procurement discrepancy but rather from disagreement with the company’s expressed views, which legal advocates contend is a form of unconstitutional punishment for speech. An amicus brief filed in the case presents “expressive governance” as the government’s method of using discretionary administrative authority—such as contracting and licensing decisions—to suppress disfavored viewpoints while maintaining an appearance of normal administrative operation.
Key Facts
The case involves a significant U.S.-based AI company engaged in federal defense contracts. The designation of the company as a “supply-chain risk” typically targets foreign adversaries but has been applied here to a U.S. entity over its policy position. The President’s directive on banning its use spans all federal agencies, effectively cutting off its government business. This method of enforcement is part of a wider pattern, cited with examples including law firms, universities, media groups, and government employees who faced sanctions for their speech. Courts have previously invalidated similar punitive actions based on administrative authority when demonstrated to be retaliatory and lacking proper legal foundation.
What This Means
The case spotlights a growing constitutional dilemma: how government agencies may utilize broad administrative discretion to penalize speech without overtly violating First Amendment protections. The concept of expressive governance highlights the risk that administrative measures—funding withdrawals, licensing restrictions, contract terminations—can serve as hidden instruments of speech suppression. For AI companies and other entities dependent on government relationships, the threat is clear: taking a public policy stance disfavored by the administration can lead to economic and reputational harm under a veneer of routine regulatory actions.
This challenges courts to develop frameworks capable of discerning when such administrative decisions constitute retaliatory silencing rather than genuine managerial discretion. The amicus brief proposes three key safeguards: requiring congressional authorization for executive actions burdening speech (a clear-statement rule); shifting the burden to the government to prove decisions would occur absent considerations of speech; and recognizing higher scrutiny when expressive intermediaries—entities shaping public information flow like AI companies—are targeted. These protections are crucial in preserving open public debate and preventing chilling effects where silence or conformity become the only safe options.
Background
“Expressive governance” builds on a broader trend where courts have struck down administrative punishments for speech when employed as instruments of retaliation rather than for legitimate regulatory purposes. The approach reflects a concern that deferring to executive discretion in contracting, funding, or clearance decisions can obscure constitutional rights violations. The current case extends this principle to AI firms whose technologies influence information processing and public discourse, raising the stakes for speech freedoms in the digital age.
What Remains Unclear
The federal court has yet to rule on the substantive constitutional claims raised in Anthropic v. US Department of War. Pending is whether it will adopt the proposed framework and impose legal constraints on how administrative powers can be wielded concerning speech-related disputes. The broader applicability of this doctrine to other industries or future administrations also remains uncertain, as does the potential for legislative action that might clarify permissible government conduct in such contexts.
What Comes Next
The court is expected to consider the arguments presented in the forthcoming ruling, which could set significant precedent on executive authority limits and First Amendment protections for AI companies and other entities dependent on federal contracts or approvals. Legal observers are watching closely as this may influence how government agencies regulate AI technology providers and manage conflicts arising from policy disagreements.
Sources
This article is based on reporting and publicly available information from the following sources:
Read more AI Regulation stories on Goka World News.