Substack Launches AI Detection Tools to Verify Human Authorship in Newsletters

Substack is introducing a new layer of transparency to its platform by integrating AI detection software, allowing readers to distinguish between human-written insights and machine-generated content. The move comes as digital platforms grapple with an influx of “AI slop”—low-quality, automated content that threatens to dilute the value of creator-led ecosystems.

The newsletter giant has partnered with Pangram, an AI writing detection startup, to provide estimates on how much of a post, note, or comment was produced by a human versus an artificial intelligence. The feature is currently available within the Substack app for any text exceeding 100 characters, offering a percentage-based breakdown of the content’s origin.

Substack CEO Chris Best framed the integration as a way to protect the “hard part” of writing—the original ideas and personal connection that define the platform’s subscription-based model. During a discussion with Pangram founder Max Spero, Best noted that while software should handle the administrative burdens of publishing, the core value of a newsletter must remain a human endeavor. This philosophy aligns with Substack’s business model, which relies on readers paying for the unique perspectives of individual writers rather than generic information.

The introduction of these tools reflects a broader industry shift toward algorithmic accountability. Major tech entities, including Meta and YouTube, have already implemented mandatory labeling for AI-generated images and videos. Similarly, the European Union’s AI Act has established legal frameworks requiring transparency for synthetic content to prevent misinformation and maintain consumer trust.

To mitigate the risk of false positives—a common issue with AI detection software—Substack is giving publishers the ability to run scans on their own drafts before they go live. Writers can also report or remove scans on their published work if they believe the software has made an error. Furthermore, the platform is introducing an optional “AI Author’s note,” where creators can proactively disclose their use of AI tools in their research or editing process.

While the detection feature could potentially expose accounts that rely heavily on automation, Substack officials clarified that the goal is not to penalize AI usage. Instead, the initiative aims to encourage a “how I make this” culture, where transparency becomes a hallmark of quality. By providing these tools, the platform seeks to reassure its user base that the newsletters they subscribe to are worth the investment of both time and money.

The rise of generative AI has presented a unique challenge for text-heavy platforms. Unlike visual media, where deepfakes are often the primary concern, the written word faces a crisis of authenticity. In an era where large language models can mimic professional prose, the ability to verify human authorship has become a critical metric for platforms that market themselves on the strength of independent journalism and personal blogging.