DHS Agents Create Fake Accounts to Infiltrate Your Private Groups

TL;DR: DHS just authorized over 6,500 agents to create fake social media accounts to infiltrate private groups, friend users under false identities, and access posts you thought were private. The policy is called "masked engagement." There's no court oversight. No congressional authorization. The only restriction? Agents shouldn't engage in "substantive engagement", a term they deliberately left undefined. If you're in any private group discussing immigration, protests, or politics, assume there's a fed reading along.

What "Masked Engagement" Actually Means

On February 13, 2026, reporter Ken Klippenstein published leaked documents revealing DHS's new "masked engagement" policy. It's exactly as bad as it sounds.^[1]

Here's what DHS agents can now do:

Create fake accounts on Facebook, Instagram, TikTok, Discord, and any other platform
Friend real users under false identities
Join closed groups without revealing they're government agents
Access private posts and photos that you shared only with friends
Use browser anonymization tools to hide their government IP addresses

A senior DHS official admitted this is "just the first step in breaching people's privacy settings in ways that they are not even aware of."^[1]

They're not even pretending it's about security. It's about access.

6,500 Agents. Zero Court Orders.

Over 6,500 field agents and intelligence operatives across DHS and ICE now have access to this capability.^[1]

The program requires no judicial oversight. No warrant. No court order. DHS granted itself this power through internal policy. Neither Congress nor any court authorized it.^[1][2]

The only stated restriction is that agents should avoid "substantive engagement." The rules deliberately don't define what that means. Friending someone? Joining their group? Commenting on their posts? Apparently all fine.^[1]

When the rules are written to be broken, they're not rules. They're cover.

Who They're Watching

The leaked documents specifically mention two categories of targets:^[1]

Pro-Palestine groups in the United States
Groups suspected to contain Mexican and Mexican American transnational criminals

Read that second one again. Not "groups run by criminals." Groups that might contain them. That could be any community group in any border state. Any mutual aid network. Any neighborhood organizing chat.

Rachel Levinson-Waldman of the Brennan Center warned this "will weaken trust in government and weaken the trust that is critical to building community."^[2]

No kidding. When anyone in your private group could be a fed, you stop talking openly. That's the point.

They've Been Doing This for Years

The Brennan Center obtained over 3,000 pages of DHS documents through a FOIA lawsuit filed in 2020. What they found: this has been going on since at least 2012.^[2]

Key findings from the documents:

At least 14 of 35 account templates allow agents to create profiles with no official DHS affiliation
CBP monitoring requires only a vague "nexus" to assigned duties
ICE's Undercover Operations Handbook remains classified
ICE ran a 2017-2019 sting operation using a fictitious university to entrap foreign students

An ICE privacy officer raised concerns about violating platform terms of service. The agency's response? Keep doing it anyway.^[2]

The new "masked engagement" policy doesn't create new surveillance. It formalizes what was already happening, and expands who can do it.

This Is the Other Half of the Strategy

Remember the DHS subpoena campaign we covered last week? Hundreds of administrative subpoenas to Google, Meta, and Reddit demanding the identities of accounts that criticized ICE?

Masked engagement is the other side of that coin. Subpoenas unmask people who post publicly. Fake accounts infiltrate the groups where people post privately.

Together, they close the loop. Public criticism gets you subpoenaed. Private discussion gets you infiltrated.

The message is clear: there is no safe place to organize online.

Congress Is Trying to Fight Back

On February 5, 2026, Senators Ed Markey (D-Mass.), Jeff Merkley (D-Ore.), Ron Wyden (D-Ore.), and Rep. Pramila Jayapal (D-Wash.) introduced the "ICE Out of Our Faces Act."^[3]

The bill would:

Ban ICE and CBP from acquiring or using facial recognition technology
Require deletion of all biometric data already collected
Allow individuals and state attorneys general to sue for violations

It's backed by the EFF, ACLU, EPIC, Fight for the Future, and Human Rights First. Co-sponsors include Senators Angela Alsobrooks (D-Md.) and Bernie Sanders (I-Vt.).^[3]

But here's the problem: the bill focuses on facial recognition, not fake accounts. Even if it passes, masked engagement would continue.

We need legislation that addresses the entire surveillance apparatus, not just one piece of it.

What You Can Do

Treat Every New Member as Potential Fed

Private groups aren't private if anyone can join. Vet new members. Require vouching from existing trusted members. Don't add strangers.

Move Off Centralized Platforms

Facebook, Instagram, and Discord hand data to government agencies. Consider Signal group chats with disappearing messages. Or Matrix servers you control.

Segment Your Groups

Don't put sensitive planning in the same chat as general announcements. If one group gets infiltrated, limit what they can access.

Assume Everything Is Logged

Even with encryption, assume screenshots happen. Don't post anything you wouldn't want on the front page of a court filing.

The Real Cost

Surveillance doesn't just catch criminals. It chills speech. It stops people from joining groups, attending protests, speaking their minds.

When 6,500 government agents can create fake identities and join your private conversations, the Fourth Amendment becomes a joke. The First Amendment follows.

DHS didn't need permission. They just did it.

That's the surveillance state in 2026.