Pangram raises $9M to out your ChatGPT ghostwriter
What happened
New York startup Pangram raised a $9M round led by Menlo Ventures, with Haystack, ScOp, Script Capital, and Cadenza also chipping in. Alongside the raise, it launched Pangram 4, a next-gen text detector, plus a new image-detection model in research preview.
Why this matters
As AI-generated slop, fake citations, and bot content flood the web, distinguishing human from machine writing is becoming a real business — not a novelty. Pangram claims over 99% accuracy, even against 'humanizer' tools designed to fool detectors, which is the kind of number VCs and worried journalists alike get excited about.
The slightly cynical read
Every AI detector claims near-perfect accuracy right up until someone posts a screenshot proving it flagged their grandma's birthday card as ChatGPT output. Betting big on a cat-and-mouse game where the 'cat' (LLMs) keeps getting smarter is a bold long-term thesis.
What to watch next
Watch whether Pangram's image detector graduates from research preview to general release, and whether institutions like arXiv start actually plugging tools like this into enforcement policies rather than just writing about them.