Substack’s new tool tells you who’s been writing their newsletters

Substack, the popular newsletter and writing platform, has rolled out a significant new feature this week: an integrated AI detection tool. Powered by AI writing detection software Pangram, this new functionality allows users within the Substack app to scan posts, comments, and replies, providing an estimate of how much of the content was authored by a human versus generated by artificial intelligence. This initiative signals a broader industry movement towards greater transparency in AI-assisted content.

The integration, which Substack announced recently, enables readers to gain insight into the creation process behind their favorite newsletters. For any content exceeding 100 characters – whether a full post, a brief note, or a comment – the tool will offer an estimated breakdown of AI versus human contribution.

Substack Embraces AI Transparency

This move by Substack positions the platform alongside a growing number of digital services that are proactively addressing the proliferation of AI-generated content. Substack CEO Chris Best framed the integration positively, calling it a “good use of AI.” He recounted his long-held philosophy for the platform: handling everything for writers “except the hard part” – the unique idea and valuable insight that only a person can provide. This tool aims to preserve the value of that human contribution.

“You have to have something — an idea that’s worth reading, that’s worth caring about, that’s worth sharing. That one thing is very hard and very valuable,” Best explained in an online chat with Pangram founder Max Spero. He added that while software should handle other aspects, the human element remains paramount for the core content.

Not a Ban, But an Encouragement for Disclosure

Substack clarified that the new tool is not intended to prohibit or penalize AI-assisted writing. Instead, its primary goal is to foster transparency and encourage creators to actively disclose their use of AI. The platform will also offer an optional “AI author’s note,” allowing writers to explain their process and the extent to which AI was utilized in their work.

Publishers on Substack will also have the ability to utilize Pangram on their own drafts before publication, providing them with an opportunity to understand their content's AI detection score. Furthermore, they can report and remove scans on their own work if they believe the detection results are inaccurate.

A Broader Industry Trend Towards Labeling AI Content

Substack's new feature is part of a wider trend across the digital landscape to label AI-generated content. Social media giants like YouTube and TikTok have already implemented systems to automatically label AI-generated photos and videos. Even within the music industry, streaming services are adapting, with platforms like Spotify updating their AI policies to label tracks and, in some cases, penalizing AI-generated music, as seen with Tidal’s crackdown on monetization for such content.

This collective shift underscores a growing industry recognition of the need for provenance and clear labeling as AI tools become increasingly sophisticated and accessible for content creation. The aim is to maintain trust and help consumers distinguish between human originality and machine-generated output.

Potential Impact on the Platform

In the short term, this transparency could present a challenge for Substack. Revealing extensive AI use in certain newsletters might potentially erode reader trust within its ecosystem of independent news and blogs, or even damage the platform's reputation for hosting high-quality, authentic content. However, Substack appears to be betting on the long-term benefits.

By actively providing readers with information about AI involvement, the feature could help prevent the platform from being overwhelmed by “AI slop” – low-quality, generic content generated without significant human input. Ultimately, this move aims to bolster reader trust, ensuring that subscribers are better informed about the nature of the content they engage with and solidifying Substack’s commitment to genuine, human-driven creativity.

FAQ

Q: What is the new Substack tool and what does it do? A: Substack has integrated AI writing detection software Pangram into its app, allowing users to scan posts, comments, and replies over 100 characters to get an estimate of how much of the content was written by a human versus an AI.

Q: Is Substack prohibiting or penalizing AI-assisted writing with this new feature? A: No, Substack clarified that the tool is not meant to prohibit or penalize AI-assisted writing. Instead, it aims to encourage transparency, allowing writers to disclose their AI usage through an optional “how I make this” statement.

Q: How does this align with broader industry trends regarding AI content? A: Substack’s move is consistent with a wider industry trend where platforms like YouTube, TikTok, and various music streaming services are implementing policies to label, and in some cases, penalize AI-generated content to ensure transparency and maintain user trust.