Substack adds an AI detector to help spot blogs written by no one

Substack is adding a tool that lets readers estimate whether posts, notes, replies, or comments were written or assisted by AI. Powered by Pangram, the scanner works on content over 100 words via a “Scan for AI text” menu option, with web and iOS support rolling out now and Android coming soon.

The feature reflects growing pressure on publishing platforms to address reader trust as generative AI floods online content. Even if AI detection remains imperfect, Substack is signaling that transparency and authenticity are becoming competitive features for creator platforms.