LinkedIn’s AI Wave: Over 40% of Long Posts Written by Bots

Scroll through your LinkedIn feed and you might feel déjà vu: articulate posts, career revelations, flawless prose. They read like a trusted colleague. Only they weren’t written by one.

Pangram, an AI-detection firm, fed nearly one million posts through its tools over a two-month window and surfaced a surprising pattern. Fast Company reported on the findings: on LinkedIn, more than 40 percent of posts longer than 250 words appear to be entirely authored by artificial intelligence. That figure is not just high. It’s staggering when you consider the platform accounted for only about one-third of the total sample yet produced almost two-thirds of the AI-identified content.

More than 40 percent of long LinkedIn posts were entirely AI-written.

Why LinkedIn leads the pack

Think about the incentives. LinkedIn rewards clarity, authority, and useful narratives that spark engagement. Those are exactly the outputs AI writing tools are trained to generate. Original posts on LinkedIn were 1.35 times more likely to be AI-generated than comments, Pangram found. Even the comments—usually considered more casual—show higher AI incidence than main posts on other networks.

X, the platform formerly known as Twitter, is also awash in machine-made longform: about 25 percent of lengthy posts were fully AI-written, and another 23 percent were produced with AI assistance. Substack, unsurprisingly perhaps, shows the lowest share among text-heavy platforms—but it is not immune. More than one in five Substack posts carried the mark of AI creation or editing.

The methodology matters: Pangram’s measurements came from content viewed through their Chrome extension across LinkedIn, X, Medium, Reddit, and Substack. That vantage point is imperfect. Detection tools have false positives and negatives. Still, when multiple signals point the same way, a pattern emerges.

So what does this mean for readers and authors? Authenticity has a new taxonomy. There is human-authored content, fully automated output, and a wide grey zone of human-AI collaboration where prompts, edits, and polishing blur authorship. For a reader, the difference matters. For brands and recruiters, it changes how we value voices and judge credibility.

Regulation and platform policy will probably follow. Platforms face a balancing act: allow tools that boost productivity and content volume, while protecting trust and transparency. Detection firms and publishers are already racing to label or flag AI-produced text, but labeling alone won’t fix the underlying incentives that encourage mass-produced, polished posts.

Writers should ask themselves a sharper question: does the AI amplify a real perspective, or does it simply manufacture one? Use the tool. But don’t outsource your point of view. Readers, meanwhile, can look for three quick signals: specificity in experience, uneven or small details that reveal lived context, and a voice that makes mistakes in ways humans do. Those are harder for models to fake consistently.

There is opportunity buried in the disruption. AI can surface good ideas faster and help people structure better arguments. It can free creators from tedious drafting. But when a platform’s feed becomes a factory of near-identical, algorithm-optimized narratives, nuance and originality pay the price.

Expect the conversation to accelerate. Detection will improve. Policies will shift. And the most valuable posts will be the ones that refuse to read like every other AI-crafted update. That alone will make them stand out.