Clearview AI, exposed by the New York Times in January 2020, scraped over 30 billion photos from Facebook, Instagram, LinkedIn, and other platforms without consent to build the world's largest facial recognition database. Over 1,000 police departments and ICE used the tool to identify people from photos. Multiple countries fined Clearview a combined ~100 million euros. The ACLU sued, and in 2022 Clearview settled, agreeing not to sell the database to private companies in the US.
“A secretive company has scraped billions of your photos from social media and built a facial recognition tool that police are using to identify anyone from a single photograph.”
From “crazy” to confirmed
The Claim Is Made
This is the moment they called it crazy.
When the New York Times published its investigation into Clearview AI in January 2020, it revealed something that should have alarmed anyone who understood its implications: a small startup had quietly assembled the world's largest facial recognition database by systematically stealing billions of photos from the internet.
The claim was straightforward but massive in scope. Clearview AI, founded by Hoan Ton-That, had scraped more than 30 billion images from Facebook, Instagram, LinkedIn, YouTube, Venmo, and countless other platforms without permission or knowledge from the users whose faces appeared in those photos. The company then built these stolen images into a database that could identify people from a single photograph with remarkable accuracy.
Initially, when confronted about the practice, Clearview defended itself using a familiar playbook. The company argued it was operating within legal bounds, claiming that scraping publicly available information from the internet was fair game under existing law. Their defense rested on the idea that because photos were posted publicly, harvesting them was no different than any other automated data collection. It was a technical argument designed to obscure an uncomfortable truth: they had built a mass surveillance infrastructure without consent.
What made this claim verifiable wasn't just the Times article. The evidence came from multiple directions. Clearview's own clients—over 1,000 police departments across America, plus Immigration and Customs Enforcement—confirmed they were actively using the tool to identify suspects and immigrants. Internal documents and company materials showed exactly what the startup was doing. The scale was undeniable: 30 billion photos represented a snapshot of a significant portion of humanity's digital life.
Get the 5 biggest receipts every week, straight to your inbox — plus an exclusive PDF: The Top 10 Conspiracy Theories Proven True in 2025-2026. No spam. No agenda. Just the papers they couldn't hide.
You just read "Clearview AI secretly scraped 30+ billion photos from social…". We send ones like this every week.
No one's said anything yet. Be the first to drop your take.
Confirmed: They Were Right
The truth comes out. Officially documented.
Confirmed: They Were Right
The truth comes out. Officially documented.
The consequences arrived swiftly. In 2021 and 2022, regulators in multiple countries took action. Italy fined Clearview 10 million euros. France imposed a 20 million euro penalty. The United Kingdom and Canada also investigated and fined the company. Collectively, Clearview faced roughly 100 million euros in regulatory penalties across multiple jurisdictions.
In the United States, the American Civil Liberties Union sued on behalf of the public. Rather than fight through litigation, Clearview settled in 2022, agreeing to significant restrictions on its business model. The company committed to stop selling its facial recognition database to private companies within the US, though it retained the ability to license the tool to government agencies and law enforcement—a compromise that satisfied few privacy advocates.
What's striking about the Clearview case is how it demonstrated a fundamental gap between what technology companies believe they can do and what society is willing to accept. The company didn't deny the scraping; it denied that scraping publicly available data was wrong. That distinction matters because it exposes how outdated our legal frameworks are when confronted with modern data collection techniques.
The case proves something people have suspected for years: large-scale facial recognition infrastructure exists, it's being used by police, and it was built using data harvested without consent. That verification has real consequences. It validates concerns about privacy erosion, algorithmic bias in law enforcement, and the concentration of surveillance power in private hands. When a company can secretly build a database of 30 billion faces before anyone notices, it suggests the safeguards protecting public privacy are fundamentally inadequate for the digital age we're actually living in.
Beat the odds
This had a 0% chance of leaking — someone talked anyway.
Conspirators
~150Network
Secret kept
0.5 years
Time to 95% exposure
500+ years
