[sigmoid10]

This is virtually impossible to build. Not just because all current "AI detector" systems are fake or outright scams with accuracy comparable to a coin-flip on frontier model output, but because even if someone did build a reliable detector and released it to the public, it could be used for adversarial training and it would become worthless pretty fast.

[linolevan]

Pangram is legit. I don't work at pangram, we integrated it in our paper website and one of the cool emergent behaviors I've seen is that on AI papers with example rollouts, it will accurately mark the paper's main text as human generated and the rollouts as AI generated.

My understanding is that they strongly believe in no false positives, so it's definitely possible to slip something by them but if it marks something as AI, it very likely is.

[dolebirchwood]

> My understanding is that they strongly believe in no false positives

Who cares what they "believe" (or, more accurately, say they believe). What are the underlying processes that actually guarantee this, and what data supports it?

[jfengel]

What is a rollout in this context?

[woadwarrior01]

> Pangram is legit.

Their 99.98% accuracy claim[1] makes me doubt that.

[1]: https://www.pangram.com/solutions/chrome-extension

[Groxx]

Rather obviously they're choosing the one that makes them look best. Another they link to¹ shows 98% for example.

Much more importantly, 9/10 dentists agree it's the best.

1: https://arxiv.org/pdf/2501.15654, linked from² https://www.pangram.com/blog/third-party-pangram-evals (the second section)

2: the third study they link there is based entirely around the assumption that Pangram is correct, and seems to have been a collaboration or something as they're included in the credits area.