Substack's AI detection lets readers scan writers' posts without notifying them

Substack's July 2026 Pangram integration lets any app user scan posts, comments, and replies over 100 characters for AI-generated content. Writers receive no notification when scanned and must actively discover and contest false positives on their own work. Writers can self-check drafts and add an optional "AI Author's note," but the reader-side scan runs without their knowledge.