A million-post study finds LinkedIn is the home of AI content
Eliza / Culture and Robotics desk
AI-generated content has a home address, and it is LinkedIn.
The study
AI detection company Pangram scanned 1,002,627 posts across LinkedIn, X, Reddit, Medium and Substack, collected since April 2026 through its Chrome extension, according to Pangram and Fast Company.
The headline finding: more than 40% of LinkedIn posts longer than 250 words were flagged as fully AI-written. LinkedIn made up about a third of everything scanned, but produced 62% of all the AI content the detector found. No other platform came close to that concentration.
The rest of the internet is not clean either
Across all platforms in the study, 25.72% of long posts were classified as fully AI-generated. Roughly one in four substantial posts you scroll past, on any of these platforms, was flagged as machine-written from start to finish.
On X, the picture was 25% fully AI plus another 23% AI-assisted. Substack came in lowest of the five platforms, and even there more than a fifth of posts got flagged.
The caveats and the response
One important limitation: the dataset reflects what the extension’s users actually saw in their feeds, not the entire internet. That means the numbers describe the content surfacing in real feeds rather than a random sample of everything published, which arguably makes them more relevant to the everyday experience of these platforms, but less generalizable as a census.
LinkedIn has responded, saying it plans to detect and downrank low-effort AI content. Whether a platform whose engagement now leans this heavily on machine-written thought leadership will aggressively suppress it is an open question, and the incentives cut both ways.
The unsettling part of the study is not any single number. It is the aggregate picture: professional social media is becoming a channel where machines write, machines rank, and increasingly, machines read. The study’s uncomfortable implication is that most readers can no longer tell the difference, and the platform’s incentives do not require them to.
Sources
Related
Ghost Font hides text in motion that humans read and AI cannot
Ghost Font, built by engineer Eric Lu, encodes words in moving dots rendered in the background color, so humans perceive the letters through motion while any single frame looks like noise to AI. Lu tested it against frontier models including Claude Fable and GPT-5.6 Sol Ultra, which struggled to decode it until told the technique, and each video also embeds a decoy message. He calls it a research experiment, not a permanent shield.
29 percent of employees admit sabotaging their company's AI
In a survey of 2,400 knowledge workers across the US, UK and Europe by Writer and Workplace Intelligence, 29% of employees admitted actively sabotaging their company's AI strategy, rising to 44% among Gen Z. In the same research, 75% of executives conceded their AI strategy is more for show than a real guide.
An adversarial wrap hid a Toyota from Flock cameras
Security researcher Bill Swearingen built noRecognition, a reinforcement learning system that generates adversarial patterns. After 31 million tests it produced patterns that defeated all 11 open source detection algorithms he tested, including those behind Flock license plate readers, Axon body cameras and Clearview AI. At Def Con 2026 he worked with Donut Media to wrap a 2009 Toyota Yaris in one of the patterns and drive it past a Flock camera undetected. He is now crowdfunding shirts and hoodies with the patterns, while keeping his strongest ones offline so surveillance vendors cannot train against them.
ANOTHER News is published by ANOTHER, an AI-native content agency. Daily coverage also runs on Instagram.