Substack is adding a new way for readers to gauge whether the words in front of them may have been produced with artificial intelligence. In a blog post published Tuesday, the company said an upcoming detection feature will review posts, notes, replies, and comments, then offer an estimate of how much of the text appears to be AI-generated or assisted by AI.
The feature is built with technology from Pangram, an AI detection firm, and is being introduced on Substack’s web platform and iOS app. Support for Android is expected “soon.” To use it, readers can open the three-dot menu in the upper-right corner of a post and select “Scan for AI text,” provided the content is longer than 100 words.
Substack co-founder and CEO Chris Best said the Pangram integration is designed to give readers more clarity as AI-written material becomes increasingly common across social platforms. “The core problem is not people using AI, or the quality of its output,” Best wrote. “The problem is when there is a mismatch between a reader’s expectation and reality, especially when they unwittingly invest their attention in something with no human thought on the other end. That’s Claudefishing.”
Alongside the AI scanning tool, Substack is giving writers a new option to describe their creative process through a “How I make this” statement. Creators will also be able to run Pangram checks on their own drafts and flag results they believe are inaccurate.
Best cautioned that “Pangram can only detect whether AI was used to make the text, not whether great human care went into creating it, nor whether AI tools were used as a source.” Still, he framed the update as a trust-building measure that could help readers decide where to spend their attention. “When readers have to wonder if what they’re reading is real, it undermines trust in authorship and threatens the livelihood of writers — including those who use AI tools thoughtfully to produce work they believe in,” Best wrote. “Platforms that reward fakeness will create a race to the bottom.”
