ICE Is Monitoring 8 Billion Social Media Posts a Day

A $5.7 million AI system. A dark web scanner. A 24/7 monitoring office with 30+ agents. Your posts are the target list.

TL;DR: ICE's Homeland Security Investigations signed a $5.7 million, five-year contract with Zignal Labs in September 2025 for an AI platform that scans 8 billion social media posts per day across 100+ languages. A separate $4.2 million contract with Fivecast gives ICE tools to monitor the dark web, marketplaces, and build "digital footprints" of individuals. And ICE is planning a $20-50 million 24/7 social media monitoring office with 30+ full-time agents who can produce dossiers on targets in 30 minutes. It's the most comprehensive social media surveillance apparatus any domestic law enforcement agency has ever built.

$5.7 Million for 8 Billion Posts a Day

In September 2025, ICE's intelligence unit awarded a five-year, $5.7 million contract through government IT middleman Carahsoft for Zignal Labs, an AI-powered social media monitoring platform [1].

Zignal doesn't just search for keywords. It uses machine learning, computer vision, and optical character recognition to analyze over 8 billion posts daily in more than 100 languages. The output: what ICE calls "curated detection feeds": essentially automated target lists for enforcement operations [1].

The platform was already in use by the U.S. Secret Service starting in 2019. It's also been deployed by the Israeli military and the Pentagon [1]. Now it's pointed at social media users inside the United States.

ICE says the system supports "real-time data analysis for criminal investigations." But there's no public documentation limiting its use to criminal targets. When you can scan 8 billion posts a day, the distinction between "investigation" and "mass surveillance" gets thin.

The Dark Web Scanner

Zignal covers social media. Fivecast covers everything else.

ICE's $4.2 million contract with Australian intelligence firm Fivecast gives agents access to ONYX, a tool that conducts "automated, continuous and targeted collection of multimedia data" from social media, news, search engines, marketplaces, and the dark web [2].

ONYX searches across hundreds of millions of dark web records (sites, pages, communications, profiles, markets, documents, and leaks) through a single interface [2].

But the part that should worry you: ONYX builds "digital footprints" from biographical data. It tracks shifts in "sentiment and emotion." It digs through past posts and social connections to identify people it determines might have "violent tendencies" or hold a grudge against the agency [2].

That's not law enforcement. That's precrime.

The 24/7 Monitoring Office

Zignal and Fivecast are just the tools. The infrastructure to use them at scale is coming next.

ICE has issued a Request for Information for a $20 to $50 million social media monitoring office staffed by at least 30 full-time analysts working around the clock. The operation would run from centers in Vermont and California, with work potentially beginning in May 2026 [3].

Private contractors would scrape publicly available data, then cross-reference it with commercial datasets from data brokers like LexisNexis Accurint and Thomson Reuters CLEAR, plus government-owned databases. Analysts would be required to produce dossiers for ICE field offices within tight deadlines, sometimes just 30 minutes for a high-priority case [3].

Those dossiers feed directly into Palantir's Investigative Case Management system, which links social media data with license plate scans, biometric records, and property information [4].

Post something on Facebook. Thirty minutes later, ICE has your address, your car, your face, and your property records in a single file.

It's Not Just About Immigration

ICE says these tools target criminal investigations. The evidence says otherwise.

Pro-Palestinian activists have been targeted and jailed by immigration authorities after being doxxed online by right-wing blacklist websites [1]. Immigration agents raided New York street vendors after a right-wing influencer posted video footage online [1]. ICE is increasingly using social media to direct enforcement strategy based on what goes viral, not what constitutes a criminal threat.

Nicole M. Bennett, a researcher at Indiana University's Center for Refugee Studies, warns that when one person is flagged, "their friends, relatives, fellow organizers or any of their acquaintances can also become subjects of scrutiny" [4].

Research backs this up. After the NSA surveillance revelations, visits to Wikipedia articles on terrorism dropped sharply. People changed their behavior because they knew someone was watching. ICE monitoring social media in real time will have the same effect, but targeted specifically at immigrant communities, activists, and anyone who posts about immigration [4].

The Full Surveillance Stack

Social media monitoring doesn't exist in isolation. It's one layer in ICE's $28.7 billion surveillance budget for 2025, nearly triple the previous year [2]. Here's what feeds into the same system:

  • Phone cracking: $11 million to Cellebrite for unlocking phones and extracting data from Signal, WhatsApp, and encrypted apps. $3 million to Magnet Forensics for Graykey. $2 million to Paragon Solutions for Graphite spyware that harvests encrypted messages without user knowledge [2].
  • Facial recognition: Mobile Fortify app scanning against 200+ million photos. $10 million to Clearview AI. BI2 Technologies for iris scanning [2].
  • License plate readers: $6 million to Thomson Reuters/Motorola Solutions. Flock Safety cameras accessed for ICE queries [2].
  • Cell-site simulators: Stingray devices in TechOps Specialty Vehicles. Over 1,885 deployments documented between 2013 and 2017 [2].
  • Data integration: Palantir's ImmigrationOS ($30 million) links IRS data, immigration records, and private databases into a single search [2].

Your social media post becomes a thread. Pull it, and ICE can unravel your phone, your face, your car, your tax records, and your physical location.

What You Can Do

Lock Down Social Media

Make accounts private. Use pseudonyms where possible. Assume anything public is being scraped by AI systems scanning 8 billion posts daily.

Secure Your Phone

Keep your phone updated. Use iPhone Lockdown Mode or Android Advanced Protection. Turn your phone off at protests and border crossings: "before first unlock" protection is the strongest state [2].

Use Encrypted Messaging

Signal remains the strongest option. Be aware that ICE has contracts for spyware (Paragon's Graphite) that targets encrypted apps, but these are expensive and typically reserved for high-value targets.

Support Legal Challenges

The EFF, ACLU, and labor unions are fighting ICE surveillance in court. The EFF's Rayhunter project helps detect cell-site simulators. Advocacy works: Milwaukee just banned police facial recognition after community pressure.

What Happens Next

The Request for Information for the 24/7 monitoring office is still in the feasibility stage. If ICE moves to active contracting, the operation could launch as early as May 2026 [3].

Meanwhile, Congress introduced the ICE Out of My Face Act on February 5 to ban ICE from using facial recognition. But no legislation currently addresses social media surveillance. The legal framework hasn't caught up to the technology ICE is already deploying.

Eight billion posts a day. Thirty-minute dossiers. Dark web scraping. Emotion tracking. Precrime scoring. This isn't a surveillance plan. It's already running.

References

  1. Truthout - ICE Just Spent Millions on a Social Media Surveillance AI Program (2025)
  2. EFF - ICE Is Going on a Surveillance Shopping Spree (January 2026)
  3. EPIC / Wired - ICE Wants to Build Out a 24/7 Social Media Surveillance Team (2025)
  4. The Conversation - Always Watching: How ICE's Plan to Monitor Social Media 24/7 Threatens Privacy (2026)