Pangram Labs has done something no other AI detector has: achieved a near-zero false positive rate. 1 in 10,000 on test documents. 1 in 100,000 on scientific papers.

When Pangram flags something as AI-generated, it is AI-generated. No more “but AI detectors also flag the Constitution!” The liar's dividend is gone: you can’t hide behind “detectors don’t work, maybe I wrote it!”

Voltaire's adage is the best way to understand Pangram’s philosophy against AI slop: perfect is the enemy of good.

Instead of trying to catch every AI text and every human text with zero errors on both sides, they chose the one battle that can be won cleanly: never punishing a human by mistake.

In exchange, their false negative rate—how much AI slop escapes—is only near-zero under lab conditions, not under real-world conditions. We just don’t know how people use AI in the wild!

That trade-off—it’s better for 10 guilty people to escape than for 1 innocent person to suffer—is what makes Pangram stand out.

I encourage you to use Pangram in your daily internet use: the only way to clean up the entire digital town is for each of us to clean the sidewalk in front of our own digital homes.

Pangram is, as I see it, the first successful offensive in humanity’s reconquest of the web.

Disclosure: not a paid sponsorship.