A former Israeli commando wants to sell America 24/7 AI threat-detection—before the next school shooting. Critics say it’s pre-crime policing in disguise.
Just hours ago, a veteran of Israel’s elite units stepped onto American soil with a pitch that could reshape policing forever. His promise? An AI system that never sleeps, scanning every corner of the internet for the next mass shooter. The catch? It might watch you too. Here’s why the debate exploded online before lunch.
The Pitch That Lit the Fuse
Cohen—an ex-special ops officer whose last name is being withheld for security reasons—stood in front of a small group of investors and police chiefs this morning. He unveiled what he calls the first AI platform built specifically for U.S. law enforcement, one that scrapes social media, forums, and even gaming lobbies around the clock.
The demo was slick. A mock shooter posted cryptic lyrics on TikTok; within seconds the system flagged the account, cross-referenced prior posts, and pinged local precincts with a risk score. The room applauded. Outside the room, phones started buzzing.
By 10:43 a.m. PDT, journalist Glenn Greenwald had already quote-tweeted the leaked pitch deck. His verdict: “Post-9/11 surveillance on steroids.” The tweet rocketed past two-thousand likes in under an hour.
How the Tech Actually Works
Think of it as Google Alerts on adrenaline. The engine ingests billions of public data points—hashtags, memes, voice chats, kill-streak videos—then layers an Israeli-grade ontology on top. That ontology, refined in counter-terror raids overseas, tags everything from gun slang to suicidal emoji chains.
Risk scores aren’t binary. They float on a color wheel: green, yellow, orange, red. A red doesn’t mean “arrest this teen”; it means “send a patrol car for a welfare check.” At least that’s the sales script.
Behind the curtain, the model updates every six hours. If a new slang term for ghost guns pops up in a Discord server in Detroit, the system learns it and pushes the tweak nationwide before dinner.
The Civil Liberties Firestorm
Greenwald wasn’t the only voice. Within minutes, civil-rights attorneys piled on. The ACLU tweeted a thread warning of “algorithmic guilt by association.” Their example: a Muslim teen who quotes violent rap lyrics could be flagged for the same reason a white gamer is ignored.
Privacy hawks raised another specter—mission creep. Today it’s school shooters, tomorrow it’s protest organizers. After all, the same data streams that spot a manifesto can also map every attendee at a Black Lives Matter rally.
Law-enforcement advocates pushed back hard. One retired sheriff posted a selfie holding his granddaughter: “If this AI saves her classroom, I’m in.” The replies under his post turned into a brawl of statistics, anecdotes, and name-calling that racked up 861 retweets before noon.
What Happens Next—and What You Can Do
Cities from Austin to Atlanta have already requested demos. Procurement budgets are being dusted off. But no contracts have been signed, which means the window for public input is still cracked open.
If you’re a parent, ask your school board if they plan to opt in. If you’re a voter, call your city council and demand a public hearing before any pilot program launches. And if you’re simply a citizen who likes privacy, remember that data once collected rarely stays in one pair of hands.
The debate isn’t going away. By the time you finish reading this, the algorithm has probably scanned another million posts. The only question left is who gets to decide what the red dots mean—and what happens when the system gets it wrong.
