// about
Internal AI is the next insider threat.
We started FakeRogue because the scariest breach isn't an outsider breaking in. It's the assistant you already trust, handing the wrong person the wrong thing.
// why we exist
The market sells comfort. We sell the truth.
Every company is racing to wire AI into the tools that hold their most sensitive data. Almost none of them know what those AIs will do when a motivated insider starts asking the wrong questions.
The AI security industry answered that fear with dashboards, frameworks, and risk scores - abstractions that feel like progress but never tell you whether your assistant will leak the board deck.
So we built the opposite. A simulated rogue employee, a growing library of real attacks, and a report that says plainly what got out. No abstraction. Just evidence you can act on.
// what we believe
Three things we won’t compromise on.
Proof over posture
We don't hand you a risk score. We hand you the data we pulled. Either your AI leaked it or it didn't.
Attack to defend
You can't secure what you've never tested under fire. So we bring the fire - safely, and on your terms.
Make you the hero
Our report exists so the person who owns the AI rollout can say 'we tested this' before anyone has to ask.
// the team
Built by someone who’s done the breaking.
Martin Gottweis
Founder · ex-Meta AI engineer
Nineteen years building software, most recently on AI systems at Meta. Martin started FakeRogue after watching companies wire sensitive data into AI tools faster than anyone could test what those tools would give away - and decided the most honest way to find out was to attack them.
// before someone else does
Let us try to rob you first.
A 30-minute demo. We'll run a real extraction against a sample environment - and show you what it would find in yours.