Substack adds AI detection tool to reveal which newsletters use artificial writing

Substack has rolled out a new capability that lets subscribers assess the degree of AI involvement in the newsletters they read. The company has integrated Pangram, an AI detection tool, directly into its platform to scan posts, notes, comments, and replies for signs of machine-generated versus human-written content.

What the feature does#

Users can now see an estimate of how much of a given piece was composed by humans and how much came from AI systems. The detection is available for any post or comment longer than 100 characters. Writers also gain the ability to add optional disclosures explaining their use of AI in the creation process and can test their own drafts before publishing to review the detection results.

The broader shift toward AI transparency#

Substack joins a wider industry movement toward marking AI-generated or AI-assisted content. YouTube now automatically labels videos made with AI, social platforms flag synthetic images and video, and music streaming services have begun identifying and in some cases limiting payouts for AI-generated tracks.

Why this matters for Substack#

The move carries potential downsides in the short run—revealing widespread AI use among newsletters could undermine reader confidence and tarnish Substack's standing as a home for original, human-driven writing. Over time, however, the tool may help the platform filter out low-effort "AI slop" and rebuild trust in its editorial ecosystem.

Substack CEO Chris Best framed the feature as a proper application of AI itself: "When I used to pitch Substack to writers, one way I would do it is […] we'll do everything for you except the hard part," he said in a conversation with Pangram founder Max Spero. The company emphasizes it is not penalizing AI-assisted work but rather encouraging transparency and helping writers document their methods.