Substack now lets readers see which newsletters are written with AI

Substack has launched a new feature that allows readers to scan posts, notes, replies, and comments on its platform to estimate how much of the content was generated by artificial intelligence. The integration with Pangram, an AI writing detection tool, is available in the Substack app for any text exceeding 100 characters.

How the AI detection feature works

The Pangram integration provides a percentage estimate of AI-written versus human-written content directly within the Substack reading experience. Writers can also run the tool on their own drafts before publishing and request removal of scans they believe are inaccurate. Substack is also introducing an optional “AI Author’s Note” that creators can use to voluntarily disclose their use of AI tools in the writing process.

Why Substack is embracing AI transparency

In the short term, the move risks exposing a significant number of Substack newsletters that rely heavily on AI generation, potentially eroding trust in the platform’s ecosystem of independent journalism and commentary. However, Substack CEO Chris Best framed the decision as a long-term investment in credibility. “Software should do everything else, but I think you do want the person to do the hard part,” Best said during an online discussion with Pangram founder Max Spero. The company emphasized that the tool is not designed to prohibit or penalize AI-assisted writing, but rather to encourage writers to explain their creative process.

Broader industry context for AI labeling

Substack joins a growing number of platforms adopting AI content labels. Social media sites now label AI-generated photos and videos, and music streaming services have begun distinguishing and sometimes penalizing AI-generated tracks. The move reflects a broader industry push to help audiences distinguish between human and machine-generated content as AI tools become more prevalent in content creation.

What this means for writers and readers

For readers, the feature offers a new layer of transparency that can inform trust decisions about the newsletters they follow. For writers, it creates an incentive to either disclose their use of AI or demonstrate their human authorship. The feature applies to posts, notes, replies, and comments, covering a wide range of interactions on the platform. Substack has clarified that the tool is meant to foster honesty rather than enforce compliance, positioning the optional author’s note as a way for creators to build credibility with their audience.

Conclusion

Substack’s integration of Pangram AI detection represents a significant step toward content transparency in the newsletter ecosystem. While it may initially expose heavy AI use among some publishers, the long-term effect could strengthen reader trust and differentiate genuinely human-crafted content in an increasingly automated media landscape.

FAQs

Q1: How does Substack’s AI detection feature work?
The feature uses Pangram’s detection software to analyze posts, notes, replies, and comments on the Substack app for text over 100 characters, providing an estimate of AI versus human authorship.

Q2: Can writers opt out of having their content scanned?
Writers can run Pangram on their own drafts before publication and can request removal of scans on their work if they believe the results are inaccurate.

Q3: Is Substack penalizing writers who use AI?
No. Substack has stated the tool is not meant to prohibit or penalize AI-assisted writing, but rather to encourage voluntary disclosure through an optional AI Author’s Note.