The State Department Is Denying Visas to Fact-Checkers

8 min read

From: Lurk More

In December 2025, the Trump administration classified content moderation as 'censorship' and instructed consular officers to deny visas to anyone who ever worked in trust and safety. The government is now formally punishing the people who tried to make the internet less terrible.
Contents 8 sections

Lurk More Newsletter

In December 2025, the U.S. State Department sent an internal memo to consular officers worldwide. The memo directed them to “thoroughly explore” the work histories of H-1B visa applicants – new and returning – by reviewing their resumes, LinkedIn profiles, and media appearances. They were told to look for evidence of involvement in content moderation, fact-checking, combating misinformation, trust and safety, or compliance work. If such evidence was found, consular officers were instructed to pursue a finding of visa ineligibility.

The rationale: these activities constitute “censorship” of protected American speech.

Read that again. The United States government has decided that the people who spent the last decade trying to remove death threats, child exploitation material, terrorist recruitment videos, and coordinated harassment campaigns from the internet were engaged in censorship. And the punishment is exile.

The Memo

The directive, first reported by NPR on December 4, 2025, targets a specific and deliberately broad category of workers. Not just people who worked at Facebook or Twitter. Anyone whose resume shows involvement in:

  • Combating misinformation, disinformation, or false narratives
  • Fact-checking
  • Content moderation
  • Compliance
  • Trust and safety

That last category is worth lingering on. “Trust and safety” is the industry term for the teams that handle everything from spam filtering to removing child sexual abuse material. It is the broadest possible description of the work that keeps platforms from becoming open sewers. The State Department has decided that all of it – every bit of it – is censorship.

The memo also arrived alongside a separate announcement requiring H-1B applicants and their dependents to set all social media profiles to “public” for review by U.S. officials. So the government that is punishing people for moderating online speech is simultaneously demanding the right to surveil everyone’s online speech. The irony is not subtle. It is not intended to be.

The visa denials are being pursued under existing ineligibility provisions of the Immigration and Nationality Act. The memo does not create new law. It reinterprets existing authority to encompass an entirely new category of undesirable: the content moderator.

This is how it works. You do not need to pass new legislation to achieve a policy goal. You send a memo. You redefine terms. “Content moderation” becomes “censorship.” “Fact-checking” becomes “suppression of protected speech.” The law does not change. The interpretation changes. And suddenly thousands of workers in a legitimate, legal, globally practiced profession are persona non grata.

Carrie DeCell, senior staff attorney at the Knight First Amendment Institute at Columbia University, called the policy “incoherent and unconstitutional,” noting that people who work on content moderation “aren’t engaged in ‘censorship’ – they’re engaged in activities that the First Amendment was designed to protect.”

She is right. She is also describing something that does not matter, because the memo is not about the First Amendment. It is about sending a message.

The Message

The message is: if you tried to make the internet less terrible, you are the enemy.

This is not new. I wrote an entire book about this pattern. In Lurk More, I describe the immune system thesis: every complex information ecosystem develops mechanisms to manage signal-to-noise ratio. Forums had moderators. Usenet had killfiles. Wikipedia has its army of volunteer editors. Social media platforms built trust and safety teams. These are the immune system of the information environment. They are imperfect, sometimes overzealous, often poorly managed. But they exist because the alternative is an information environment that becomes unusable.

The pattern – and it is always the same pattern – is that the people who benefit most from the degradation of information quality attack the immune system first. You do not need to win the argument if you can destroy the mechanism that distinguishes arguments from noise. You do not need to produce better information if you can make it impossible to distinguish good information from bad.

The content moderation industry is worth $12.48 billion as of 2025, projected to reach $42 billion by 2035. That is the scale of the problem. That is how much money it costs, globally, to keep the internet from drowning in its own effluent. The State Department has decided that this entire industry – every human being working in it who needs a visa – is engaged in censorship.

The Compliance Theater Connection

Here is where it gets interesting, if you are the kind of person who tracks institutional failure patterns across domains.

The same government that is punishing content moderators for “censorship” spent the last twenty years building the largest compliance apparatus in history. PCI-DSS. SOX. HIPAA. GDPR enforcement cooperation. The entire regulatory state is built on the premise that institutions need oversight mechanisms, that self-regulation requires verification, that trust must be earned through demonstrated compliance.

Content moderation is compliance. It is the platform equivalent of the audit function. When Facebook removes a post that violates its terms of service, it is doing exactly what every compliance framework in every industry requires: enforcing the rules it agreed to follow. When a fact-checker flags a false claim, they are performing the verification function that every quality-assurance regime demands.

The government is not attacking censorship. The government is attacking compliance. It is punishing the auditors.

I have seen this before. In Lurk More, I trace the arc from PCI compliance theater through cyber insurance to AI safety kabuki. The pattern is always the same: build the compliance mechanism, watch it fail to prevent the thing it was designed to prevent, then blame the mechanism instead of the underlying pathology. The content moderation version is: build the trust and safety team, watch it fail to perfectly moderate billions of posts per day, then punish the team for trying.

Who This Actually Hurts

The visa restrictions target H-1B holders. These are, overwhelmingly, skilled immigrants working in technology. Many of them are from India, China, and other countries with large tech workforces. The people being punished are not American citizens making domestic policy decisions. They are foreign workers who took jobs at American companies doing work those companies deemed necessary.

Some of these people have been in the United States for years. They have families, mortgages, children in American schools. Their visa renewals now depend on whether a consular officer decides that their job title – “Trust and Safety Analyst,” “Content Policy Specialist,” “Integrity Engineer” – constitutes evidence of censorship.

The chilling effect is the point. You do not need to actually deny thousands of visas. You need to make the category toxic. You need every tech company to think twice before putting “content moderation” in a job description. You need every foreign worker to scrub their LinkedIn of any reference to trust and safety work. You need the immune system to dismantle itself.

The Precedent

There is no precedent for a government classifying a private-sector compliance function as grounds for visa denial. Governments have denied visas to spies, to criminals, to terrorists, to people with communicable diseases. The United States has, at various points in its history, denied visas to communists, to anarchists, to people who advocated the overthrow of the government.

It has never before denied visas to people whose job was removing spam.

That is where we are. The person who spent eight years at a social media company making sure that scam ads, bot networks, and harassment campaigns were removed from the platform is now, in the eyes of the United States government, in the same category as someone who poses a threat to national security.

The internet did not get better when the moderators left. It will not get better when the government punishes the ones who remain.

The Thesis

In Lurk More, I argue that trolling – real trolling, not the word the media uses to mean “being mean online” – is the immune system testing itself. The troll probes the boundary. The moderator enforces the boundary. The tension between them is what keeps the system honest. Remove the troll and the system becomes brittle. Remove the moderator and the system becomes unusable.

The State Department has chosen a side. It has chosen the side of noise over signal, of entropy over order, of the open sewer over the imperfect filter.

The people who built those filters were not heroes. Many of them were underpaid contractors in the Philippines reviewing beheading videos for eight hours a day. Some of them developed PTSD. The industry that employed them was often exploitative, poorly regulated, and inadequately funded. There is plenty to criticize about how content moderation has been practiced.

But the answer to “content moderation is imperfect” is not “content moderation is censorship.” The answer to “the immune system sometimes attacks healthy tissue” is not “destroy the immune system.”

We know what happens when you destroy the immune system. The patient does not become free. The patient becomes sick.

This essay draws from Lurk More, coming fall 2026.

Sources