Substack Integrates AI Detection to Label AI-Written Newsletters

Substack adds AI detection scanning to its platform

Substack has integrated AI detection software into its newsletter platform, giving readers the ability to check whether their favorite writers are using artificial intelligence to generate content. The company announced this week that it has partnered with Pangram, an AI writing detection tool, to scan posts, comments, and replies across the platform.

The feature displays an estimate of how much content was written by a human versus generated by AI. Any text over 100 characters on the Substack app can be scanned, providing readers with transparency about the authorship of what they're consuming.

Why it matters

This move represents a calculated risk for Substack's business model. The platform has built its reputation on hosting independent writers and high-quality journalism, but widespread AI usage could undermine reader trust. By proactively offering detection tools, Substack is betting that transparency will ultimately strengthen its ecosystem more than concealment would. The decision also positions the company alongside other platforms like social media sites and music streaming services that have begun labeling AI-generated content, establishing transparency as an industry standard rather than an exception.

Transparency over prohibition

Substack CEO Chris Best emphasized that the tool is designed to inform rather than punish. "This is good use of AI," Best said during an online conversation with Pangram founder Max Spero. He explained that Substack's philosophy has always been to handle the technical infrastructure while writers focus on the creative work.

"You have to have something — an idea that's worth reading, that's worth caring about, that's worth sharing," Best said. "That one thing is very hard and very valuable... software should do everything else, but I think you do want the person to do the hard part."

The platform will allow writers to add an optional AI Author's note where they can disclose their use of AI tools. Substack characterized this as encouraging a "how I make this" statement rather than implementing penalties for AI-assisted writing.

Writers can also scan their own drafts before publication and request removal of scans they believe are inaccurate.

Short-term risk, long-term strategy

The integration could expose newsletters that rely heavily on AI generation, potentially damaging trust in individual publications or the platform itself. However, Substack appears to be prioritizing long-term platform integrity over short-term revenue protection.

By implementing detection features now, the company aims to prevent an accumulation of what it calls "AI slop" — low-quality, automated content that could degrade the overall user experience. The approach mirrors broader industry trends, as platforms across social media, photography, and music have begun labeling AI-generated content to maintain user trust.

The details were first reported by TechCrunch.