A federal hearing over Anthropic’s access to government work shifted toward a constitutional issue: how the administration explains a national-security label when political speech is in the record.
A lawsuit over the Trump administration’s treatment of Anthropic is now focused not only on AI and military procurement, but on what the government may do when a contractor publicly disagrees with officials.
According to The New Republic’s report on a Thursday hearing, U.S. District Judge Rita Lin pressed Justice Department lawyers on that point and was not satisfied with their response. The exchange came as the government defended the Defense Department’s decision to classify Anthropic as a national security risk and President Donald Trump’s order barring federal agencies from using the company’s services.
The contract question became the flashpoint
Lin’s most pointed hypothetical involved defense contracts and criticism of Trump.
The New Republic reported that Lin asked whether it would be unconstitutional for the government to terminate every defense contract held by anyone who had ever publicly criticized Trump. DOJ lawyers hedged rather than giving the direct answer she appeared to expect.
“I’m surprised that you can’t give a yes to what seems to be an easy question,” Lin said, according to the report.
The moment mattered because it framed the dispute in First Amendment terms. The question was not whether the government has discretion in defense contracting; it was whether that discretion could extend to punishment for protected political speech.
Lin had previewed the issues
The hesitation was especially notable because, according to The New Republic, Lin had issued a legal notice earlier in the week identifying questions she planned to ask the government’s attorneys.
At the outset of the hearing, Lin said the record had “gotten worse” for the government, according to The New Republic’s account, which cited the Law & Crime network.
She also said the record appeared “largely the same” as what she had seen during the preliminary injunction stage and questioned whether the government had provided evidence “really justifying what it did,” according to the report.
Those comments do not decide the case. They do, however, indicate that Lin was looking for support in the record for the government’s national-security rationale and did not see enough in the public account of the hearing.
How Anthropic ended up in court
Anthropic, the company behind Claude, entered the dispute after it refused to grant the Pentagon unrestricted use of its chatbot for military purposes, according to The New Republic.
The report said Anthropic argued Claude could not safely or reliably be used for “autonomous lethal warfare and mass surveillance of Americans.”
Trump and Defense Secretary Pete Hegseth later said Anthropic had gone “WOKE,” The New Republic reported. The Defense Department then treated Anthropic as a national security risk, and Trump’s order barred federal agencies from using its services.
Lin had criticized the administration’s approach earlier in the case, calling aspects of the effort to cut ties with the AI contractor “Orwellian,” according to the same report.
The legal pressure point
The government can argue that courts should give the executive branch room on national-security and procurement decisions. Anthropic can argue that a national-security label cannot be used to retaliate against a company for its stated objections to certain military AI uses.
Lin’s questions appeared to test that boundary. If the government’s position depends on a concrete security justification, the record becomes central. If the action was tied to disfavored speech or viewpoint, the First Amendment concerns become harder to avoid.
The New Republic reported that Lin found the administration’s position “really troubling” in relation to the First Amendment, particularly where contracts involving those who might do business with a labeled target could be affected.
That is why the DOJ’s answer to the Trump-criticism hypothetical carried weight. A narrow defense of a specific security decision is different from a theory that could reach contractors because of political criticism.
What remains unresolved
The hearing took place during cross motions for summary judgment, according to The New Republic’s excerpt, meaning both sides are asking the court to rule on the existing legal record.
Lin has not ruled based on the comments reported from the hearing. The next major development will be whether she upholds, narrows or blocks the administration’s action against Anthropic.
The case remains a test of how the government supports a national-security claim when the record also includes political language about a contractor. For Anthropic and other AI firms that may contract with federal agencies, the issue is whether objections to military uses of AI can be separated from eligibility for government work.
For the Justice Department, Lin’s reported questions show the burden ahead: explain the administration’s action with evidence sufficient to withstand constitutional scrutiny, not just with the label of national security.
Leave a Reply