Substack Now Lets Readers Scan Posts for AI-Written Text
A new detection tool built into Substack will estimate how much of any post was written by AI, giving readers a way to check before they invest their time.
Key points
- Substack began rolling out an AI-text detection tool on its web platform and iOS app on Tuesday.
- The tool is powered by Pangram, an AI detection company, and works on any content longer than 100 words.
- Writers can also scan their own drafts with Pangram before publishing and flag results they believe are wrong.
- Substack CEO Chris Best coined the term "Claudefishing" to describe content that tricks readers into thinking a human wrote it.
Substack wants its readers to know whether a real person wrote what they are reading, or whether a machine did most of the work.
The newsletter and social platform has started rolling out a built-in AI detection tool, powered by a company called Pangram, which specialises in identifying AI-generated writing. Readers can trigger a scan on any post, note, reply, or comment that is longer than 100 words. Just tap the three-dot menu in the top-right corner of a post and choose "Scan for AI text." The tool gives back an estimate of how much of the writing could have come from an AI, or been heavily assisted by one.
The feature is live on the web and the Substack iOS app right now. Android support is coming soon, first reported by The Verge AI.
Substack co-founder and CEO Chris Best framed the move as a trust problem, not a technology ban. "The core problem is not people using AI, or the quality of its output," Best wrote in a blog post. "The problem is when there is a mismatch between a reader's expectation and reality."
Best gave that mismatch a name: Claudefishing. It is a play on catfishing, where someone pretends online to be someone they are not, except here a writer pretends human thought went into text that a large language model, the technology behind chatbots like ChatGPT, actually produced.
The tool has real limits. Best is open about them. Pangram can tell you whether AI shaped the words on the page. It cannot tell you whether the writer spent hours carefully directing, editing, and fact-checking that AI output. A thoughtfully assisted piece and a lazily generated one could score similarly.
To handle that gap, Substack is also adding a voluntary "How I make this" statement, where writers can explain their process in their own words. Think of it as a nutrition label for creative work.
Should readers trust the scan results?
Treat them as a useful signal, not a verdict. Pangram's estimate is a probability, not a proof. Best himself notes it can produce inaccurate results, which is exactly why writers get an option to flag and report a score they think is wrong.
For everyday Substack readers, the practical upshot is simple. If you are paying for a newsletter or giving a writer your attention, you now have a quick way to sense-check whether a human actually crafted what you are reading. That is more than most platforms offer.
Best put it plainly: "Platforms that reward fakeness will create a race to the bottom."
