Pangram Raises Funding to Detect AI Content
Startup releases new text and image detection models as AI-generated content proliferates online.
TLDR
Pangram, developing tools to identify AI-generated writing and imagery, has raised funding to scale its software. The company released Pangram 4 for text detection and an image detection model in research preview. Backed by Menlo Ventures, it claims top performance in independent academic tests with a low false-positive rate. Coverage from TechCrunch and The New York Times frames the effort as a response to AI slop flooding the internet, raising questions about accuracy and use cases for such detection.
Combined views
6.2K
4 Sources, first seen ago
Sources
K- 
K-A-T-U-R-E-B-E@Katurebe99Pangram is a very weak product for the following reasons: 1) The idea of fighting writing with AI is akin to fighting typing in favor of writing with hands, for the tool is a multiplier not the primary originator of output/content. 2) LLMs are trained on decades-old human…
SU sunmer@sunmer575399Tried a few detection tools last month on my own writing. They flagged 70% of my human-written blog posts as AI. Meanwhile actual GPT output slipped through half the time. Pangram's $9M might buy good marketing but detecting AI text reliably is still a myth. I'd rather see them…
NA Numana Afroje Opy@dorryssss@TechCrunch A much-needed initiative! As deepfakes and AI spam increase, spotting the difference is critical. Raising $9M shows just how vital content verification is right now. Best of luck to the Pangram team! 🛡️💻
MS Max Spero@max_spero_@deedydas It's been great working with you so far, Deedy, and I'm so excited for what's to come!
AL alice@aliceisplaying@TechCrunch hell yeah