Substack has launched a detection tool that estimates the percentage of AI-generated content in newsletters published on its platform. The feature appears directly on newsletter pages, helping readers identify which pieces relied heavily on artificial intelligence during the writing process.
The tool reflects growing pressure across publishing platforms to disclose AI involvement in content creation. As AI writing assistants become more sophisticated and widely adopted, readers increasingly want visibility into whether they're engaging with human-authored work or machine-assisted text.
Substack's approach differs from outright bans on AI content. Instead, the platform opts for transparency, allowing creators to continue using AI while making that use visible. This middle-ground strategy acknowledges that AI tools can assist legitimate editorial work—research summaries, drafting, editing—without requiring creators to abandon them entirely.
The detection mechanism likely relies on statistical analysis of text patterns, word choice frequencies, and syntactic structures that distinguish AI-generated language from human writing. Such tools remain imperfect. They can flag human-written content as AI-generated and miss genuinely machine-written sections, but they offer readers a useful heuristic.
For newsletter creators, the tool creates an incentive structure. Writers who want to build trust with subscribers have reason to minimize AI detection scores or clearly label AI assistance themselves. Creators relying entirely on AI generation risk losing audience confidence.
This move aligns with broader industry movements toward AI disclosure. Publications like The New York Times and NPR have established policies requiring disclosure when AI assisted in reporting or writing. Search platforms and social media companies face similar pressures from regulators and users demanding clarity around AI involvement.
Substack's transparent approach differs from platforms that attempt to completely eliminate AI-generated content, which creates enforcement problems and limits creator flexibility. By making AI use visible rather than invisible, Substack lets readers make informed choices about what they read.
The long-term impact depends on user behavior.