Mark RadarMARK RADAR
About
EN
Sign in
Event File AI Generative AI

Pangram CEO Warns AI Content Is Pushing Web Toward ‘Dead Internet’

2 reports · First detected 2026-09-03 · Last active 2026-09-03

The spread of cheap, machine-generated text, images and online interactions is eroding the distinction between human and automated activity. Max Spero, chief executive of AI-detection startup Pangram, says that loss of provenance threatens trust across the web and is pushing it “dangerously close” to the so-called dead internet theory — the idea that bots and synthetic material will dominate online discourse.

Spero said in recent interviews that detection is more complicated than labeling material simply real or fake, because content can combine human and machine edits while generative models keep changing. Pangram has raised $9 million and is working with publishing platform Substack to develop tools aimed at identifying AI-generated content more accurately. The reports did not specify exact dates for the funding round or partnership.

All Coverage

2 original reports

The Backstory

The history behind this event
Pangram Raises $9 Million to Expand AI Content Detection2026-07-29 · 1 reports · similarity 0.85

Generative AI is spreading across news, academic papers, social media and search-driven content, increasing demand for tools that distinguish human work from machine-generated or AI-edited material. New York-based Pangram was founded about two years ago by Stanford AI and machine-learning graduates Max Spero and Bradley Emi. Its detection system compares human documents with synthetic counterparts to identify stylistic patterns, targeting publishers, schools, online platforms and other organizations seeking to preserve trust in digital content.

On July 29, 2026, Pangram said it raised $9 million in a round led by Menlo Ventures, with Haystack, ScOp, Script Capital and Cadenza participating. The company simultaneously launched Pangram 4, which it says is more than 99% accurate at identifying AI-assisted and mixed human-AI writing, and unveiled Pangram Image in research preview ahead of a broader release in coming weeks. Pangram says about one in 10,000 human documents is wrongly classified as AI, while customers including Substack and Quora access its technology through an API.

Mark Radar|MARK RADAR

If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →

All times are in Taipei time (GMT+8)