How and Why the Government is Stalking You
Government surveillance in the AI era — from CIA social media monitoring to modern AI-powered surveillance.
title: "How and Why the Government is Stalking You" slug: "how-and-why-government-stalking-you" description: "Government surveillance in the AI era — from CIA social media monitoring to modern AI-powered surveillance." datePublished: "2013-03-18" dateModified: "2026-03-15" category: "Government & Data" tags: ["surveillance", "government", "privacy", "social media"] tier: 1 originalUrl: "http://www.applieddatalabs.com/content/how-and-why-government-stalking-you" waybackUrl: "https://web.archive.org/web/20130318051333/http://www.applieddatalabs.com:80/content/how-and-why-government-stalking-you"
How and Why the Government is Stalking You
We published this piece in early 2013, months before Edward Snowden walked out of an NSA facility in Hawaii with a thumb drive that would change the world. At the time, we thought we were being provocative. We wrote about the CIA monitoring Twitter and Facebook, about a defense contractor called Raytheon building a tool that could predict where you'd be next Tuesday, and about a government program running 500 fake social media accounts to manipulate online conversations. Readers probably thought we were being a little paranoid. Turns out we were being optimistic about how restrained the surveillance actually was.
What We Knew in 2013
Our original article opened with the Arab Spring, where protesters famously said, "We use Facebook to schedule the protests, Twitter to coordinate, and YouTube to tell the world." We argued that governments weren't cracking down on social media. They were leaning into it as an intelligence goldmine.
We documented three specific programs. First, In-Q-Tel, the CIA's venture capital arm, had invested in a company called Visible Technologies that crawled over half a million web 2.0 sites per day, scraping more than a million posts from blogs, forums, Flickr, YouTube, Twitter, and Amazon. A Visible spokesman told reporters the CIA wanted "early-warning detection on how issues are playing internationally."
Second, Raytheon had built a tool called RIOT (Rapid Information Overlay Technology) that aggregated social media data to build psychological profiles, map social networks, and predict a person's future location based on their check-in patterns.
Third, and most disturbing, we reported on Operation Earnest Voice, a program run by U.S. Central Command through a California company called Ntrepid. The system let 50 real operators manage 500 fake personas, each "replete with background, history, supporting details, and cyber presences that are technically, culturally and geographically consistent." The stated purpose was countering extremist propaganda on foreign-language websites. But as RAND Corporation researcher Isaac R. Porche pointed out at the time, it's not easy to exclude U.S. audiences when you're operating on the open internet.
We wrote this article months before Snowden. We thought we were being provocative. We were actually underestimating the scope of what was happening by several orders of magnitude.
Then Snowden Happened
In June 2013, three months after our article, Edward Snowden leaked thousands of classified NSA documents to journalists Glenn Greenwald and Laura Poitras. The revelations were breathtaking. PRISM gave the NSA direct access to servers at Google, Facebook, Apple, Microsoft, Yahoo, and other major tech companies. The agency was collecting phone metadata on virtually every American through a secret court order to Verizon. XKeyscore let analysts search through vast databases of emails, chats, and browsing histories with no warrant required. Upstream collection tapped directly into the fiber optic cables carrying internet traffic.
Everything we'd written about in 2013 was, in retrospect, the tip of the iceberg. The Visible Technologies contract and RIOT prototype were quaint compared to the NSA's industrial-scale surveillance infrastructure. The government wasn't just monitoring social media. It had built a system to capture and store essentially all digital communications.
The fallout reshaped the tech industry. Apple began encrypting iPhones by default in 2014. WhatsApp rolled out end-to-end encryption in 2016. The EU passed the General Data Protection Regulation (GDPR) in 2016, effective 2018, partly in response to revelations about NSA surveillance of European citizens and leaders.
AI Made Everything Worse
If the Snowden era was about bulk data collection, the 2020s are about what you can do with all that data once you add AI.
Clearview AI scraped over 30 billion photos from social media and the open web to build a facial recognition database that it sold to more than 600 law enforcement agencies in the U.S. alone. In 2022, Ukraine's government used Clearview AI to identify Russian soldiers and dead combatants on the battlefield. The technology went from controversial startup to wartime tool in under three years.
Palantir, co-founded by Peter Thiel with early CIA funding through In-Q-Tel (the same outfit we wrote about in 2013), landed a $463 million contract with the U.S. Army in 2024 for its Maven Smart System. The company's Gotham platform is used by intelligence agencies, immigration enforcement (ICE has been a major customer since 2014), and police departments. Palantir's market cap crossed $150 billion in early 2025, making AI-powered surveillance one of the most valuable business models in America.
Predictive policing tools like PredPol (now rebranded as Geolitica, then quietly shut down in 2023 after years of criticism) and ShotSpotter spread through hundreds of police departments before studies showed they disproportionately targeted Black and Latino neighborhoods without reducing crime. Chicago, Los Angeles, and New Orleans all adopted and then scaled back these systems under public pressure.
Meanwhile, China built the world's most comprehensive AI surveillance state, with over 600 million cameras, mandatory facial recognition at transit hubs, and social credit scoring. The technology stack is built by companies like Hikvision, Dahua, and SenseTime, several of which were sanctioned by the U.S. government even as American law enforcement agencies purchased similar capabilities from domestic vendors.
What Enterprises Need to Understand
For businesses, operating in a surveillance-aware world means rethinking how you handle data at a fundamental level. If you're collecting customer location data, biometric data, or behavioral data, you need to assume that government agencies may seek access to it. This isn't theoretical. In 2018, it came out that the FBI was buying Americans' location data from commercial data brokers, sidestepping the warrant requirement entirely. In 2023, the FTC started cracking down on data brokers selling sensitive location data, fining Kochava and X-Mode Social (now Outlogic) for tracking visits to sensitive locations like reproductive health clinics and places of worship.
Enterprise AI systems that process personal data need robust governance frameworks not just for compliance, but because the data you collect can end up in places you never intended. Building AI-ready data infrastructure means designing for privacy from the start, not bolting it on after a breach or subpoena makes the news. The Operational AI approach treats data privacy as an operational discipline, not a one-time legal review.
Where This Leaves Us
I wrote back in 2013 that social networks offer a trade-off, and I asked readers to think before posting: "Do I want everyone to know this about me?" That advice sounds almost naive now. The question in 2026 isn't whether you're comfortable sharing. It's whether you have any meaningful choice. Your face is in Clearview's database whether you posted it or not. Your location history exists in a dozen data brokers' systems through your phone's advertising ID. The RIOT prototype that scared us in 2013 would be a toy compared to what a mid-tier police department can buy off the shelf today.
The surveillance tools we warned about didn't go away. They got an AI upgrade, went commercial, and scaled to every law enforcement agency with a budget line for software subscriptions.