Anthropic is battling the federal government in two lawsuits—one in San Francisco, the other in Washington, DC. Both cases center on Secretary of War Pete Hegseth designating the artificial intelligence company a supply-chain risk to national security and terminating its business with the Department of War (DoW) after the government transitions to what Hegseth called “a better and more patriotic service.”

The lawsuits include a First Amendment retaliation claim against the DoW and Hegseth. I’ve explained that it pivots on the premise that the defendants punished Anthropic “for publicly expressing—and sticking to—its safety, risk, and civil liberties-grounded viewpoint that the DoW not deploy Anthropic’s Claude Gov model for ‘mass domestic surveillance of Americans and fully autonomous weapons.’”

Hegseth, in contrast, wanted to implement Anthropic’s AI tools for “all lawful uses.” He scoffed at the company’s stance against mass domestic surveillance, deriding it as “corporate virtue-signaling that places Silicon Valley ideology above American lives.” Hegseth’s statement followed President Donald Trump’s Truth Social post lambasting Anthropic as an “out-of-control, Radical Left AI company run by people who have no idea what the real World is all about.”

However, if the results of an April poll of 1,000 Americans conducted for the Foundation for Individual Rights and Expression (FIRE) are accurate, then many real-world people seemingly understand Anthropic’s position. Specifically, 77 percent of respondents were “somewhat,” “very,” or “extremely” concerned about the federal government “pressuring tech companies to provide it with unrestricted access to their products so they can be used for mass surveillance.” That’s more than three times the percentage of Americans who said they were either “not at all” or only “slightly” concerned about such government pressure.

It isn’t the only area where Americans seemingly are bothered by jawboning—informal censorship that occurs when the government pressures private businesses to change their own views or to squelch or deprioritize viewpoints and messages of others that the government disdains. Jawboning violates the First Amendment when government pressure pushes past the bounds of permissible persuasion and enters the realm of illicit coercion, typically via threats of adverse consequences if businesses fail to follow the government’s censorial desires.

The Supreme Court recently explained that “[g]overnment officials cannot attempt to coerce private parties in order to punish or suppress views that the government disfavors.” Speech intermediaries—platforms such as Facebook and X that host others’ content—thus can find themselves targets of jawboning by government officials to get them to stifle content that others post. The Court has determined that:

To state a claim that the government violated the First Amendment through coercion of a third party, a plaintiff must plausibly allege conduct that, viewed in context, could be reasonably understood to convey a threat of adverse government action in order to punish or suppress the plaintiff’s speech.

For instance, I recently addressed a federal judge’s April ruling in Rosado v. Bondi that two former, high-ranking Trump administration officials likely violated the First Amendment rights of a woman by coercing Facebook to remove her “ICE Sightings—Chicagoland” group and by coercing Apple to jettison from its App Store an application called “Eyes Up.” As Judge Jorge L. Alonso explained, both of the government’s targets “allow users to post videos and information regarding ICE activity,” and the Trump administration objects to publishing such lawful content. Alonso’s May 5 preliminary injunction bars the government from taking “formal or informal” actions “to demand or pressure” Facebook and Apple “to suppress or continue to suppress” the targeted speech “or to coerce or threaten [the companies] with sanctions for allowing or reinstating” the “ICE Sightings—Chicagoland” group or “Eyes Up” application. In brief, Alonso’s order pushes back against allegedly illicit jawboning.

Americans may be tracking jawboning cases such as Rosado v. Bondi. FIRE’s poll found that 75 percent of respondents were “somewhat,” “very,” or “extremely” concerned about the federal government “pressuring social media companies to suppress the posting of certain viewpoints.” Only 25 percent were “not at all” or just “slightly” concerned about such pressure on social media companies. The high level of public concern about government pressure on social media platforms is particularly remarkable given the criticism, lawsuits, and negative publicity platforms currently face for supposedly addicting and harming minors.

Similarly high levels of public concern also exist about the federal government “pressuring media organizations to report news about foreign affairs in a favorable way” and “pressuring video platforms to remove content because of the viewpoints expressed.”

People are worried, as the stench of government jawboning fills their nostrils during an election year. Compounding the problem, 58 percent of FIRE’s respondents expressed either “no confidence at all” or “very little confidence” that Trump will protect their First Amendment rights. The US Constitution and the First Amendment thus might just matter to some voters come November.