The modern inbox has become a battlefield of efficiency. For the independent writer, the pressure to maintain a consistent publishing cadence has led to a quiet, widespread adoption of large language models to draft outlines, polish prose, and generate summaries. However, as the volume of AI-assisted content surges, a subtle tension has emerged between the creator and the reader. The audience can often sense the sterile cadence of a machine, leading to a creeping erosion of trust in the very intimacy that makes the newsletter format successful. The industry is currently grappling with a fundamental question: does the use of AI diminish the value of a human perspective, or is it simply a new tool for the modern scribe?
The Mechanics of Content Transparency
Substack is addressing this tension by integrating Pangram, an AI detection software designed to estimate the balance between human and machine authorship. This integration allows for the immediate display of estimated human-to-AI writing ratios directly within the app. The scope of this monitoring is comprehensive, extending beyond the primary newsletter posts to include notes, comments, and replies. Specifically, the detection system triggers for any piece of text content that exceeds 100 characters, ensuring that even short-form interactions are subject to the same transparency standards.
Rather than acting as a policing mechanism to ban or penalize AI usage, the tool is positioned as a transparency layer. Substack is encouraging writers to be proactive about their process through a new optional AI author note feature. This allows creators to include a how I make this statement, where they can explicitly describe their workflow and the role AI played in the creation of the piece. To prevent the frustration of false positives, publishers have the ability to pre-scan their drafts using Pangram before hitting publish. If a scan produces a result that the author believes is inaccurate, they can report the discrepancy and request the removal of the scan result.
The War Against AI Slop
This move is not an isolated experiment but part of a broader global shift toward labeling synthetic media. We have already seen social media giants implement labels for AI-generated images and videos, and music streaming platforms have begun flagging AI-generated audio, with some even applying penalties to synthetic tracks to protect human artists. Substack is now extending this logic to the realm of long-form text. The core objective is to combat the rise of AI slop—the deluge of low-effort, machine-generated content that threatens to drown out genuine intellectual contribution.
The philosophical divide here is clear. In a recent exchange with Pangram founder Max Spero, Substack CEO Chris Best emphasized that while software can handle the mechanical aspects of writing, it cannot replace the generation of a truly valuable idea. Best has long maintained that the platform's role is to provide every tool necessary for the writer, except for the one thing that cannot be automated: the ability to produce ideas that people actually care about and want to share. By surfacing the AI ratio, Substack is effectively shifting the value proposition of a newsletter from the polished final product to the intellectual labor behind it.
There is a significant short-term risk inherent in this strategy. By exposing the AI ratios of existing newsletters, Substack may inadvertently reveal that a large portion of its hosted content is not entirely human-authored. This could lead to a temporary dip in the perceived prestige of the platform as the reality of AI dependency becomes visible. However, the long-term play is a strategic purge. By making the distinction between human and machine authorship explicit, Substack protects the trust assets of its most dedicated independent writers, ensuring that their human-led insights are not diluted by a sea of synthetic approximations.
Transparency regarding the creative process is becoming the primary metric for trust in the generative era.
