Substack rolled out new tools to detect AI-generated content and stirred a hornet’s nest. The company aims to spotlight AI use in newsletters but met fierce pushback.
CEO Christ Best announced the move on July 20, 2026, citing a blurry line between human and machine writing. “It’s getting harder to tell what’s real on the internet,” Best said. He warned AI content was “polluting the commons.”
Substack’s tool, powered by Pangram, assigns a percentage score estimating how much writing is AI or human-generated. Pangram’s CEO Max Spero claims a false positive rate near one in 10,000. Still, creators reject the system as a “witch hunt.”
Some AI users say the policy unfairly singles them out. Others fear their genuine work might get flagged mistakenly. Ghostwriter Alice Lemee highlighted the stakes: “All it takes is one false accusation for a writer to have their reputation almost irreversibly tarnished.”
Substack tries to balance these concerns. They introduced a “How I make this” option letting authors explain their process. The company insists the detection tools promote transparency, not censorship. A spokesperson said the features “do not impact discovery on the platform.”
The AI detection function works on posts, notes, replies, and comments longer than 100 characters. It’s available on web and iOS, with Android coming soon. Creators can scan drafts before publishing and flag incorrect scans for review.
Sam Illingworth, a professor and Slow AI author, called the approach a machine deciding if a real human wrote the text. “We have now built the mirror image,” he said. “Pangram is a real machine, trained on vast amounts of text to make a probabilistic guess.”
Illingworth praised the transparency feature, saying, “I actually really like the new feature that lets authors tell their readers how they construct their newsletters using AI.” He urged Substack to encourage dialogue instead of relying solely on imperfect scores.
Meanwhile, other platforms also grapple with AI content. YouTube rewards AI-enhanced storytelling and clarified its “inauthentic content” rules. TikTok requires AI labels and adds invisible watermarks to some videos. Pinterest started AI labels in 2025 and lets users limit AI content in feeds.
Meta has used AI info labels since June 2024. LinkedIn estimates 41% of long-form and 30% of short-form posts between April and June 2026 were AI-generated. Pangram’s data shows 21% of YouTube’s feed contains AI videos.
Substack’s new tools fit this broader trend of AI transparency policies. Christ Best said, “We’re not against people using AI to assist their work, and we think people should be free to choose which tools they use to express themselves.” He added, “People should know what they’re getting.”
Still, resistance runs deep. Influencer Mack Collier refused to apologize for AI use, saying, “I wrote for 20 years without AI, I could do it again if I wanted to.” The debate over AI’s role in writing is far from settled.
Based on
- Substackers Say New AI Detection Tool Is a ‘Witch Hunt’ — 404media.co
- Substack Is Adding An AI Detection Feature — engadget.com
- Substack’s Anti-Slop Plan Points to the Next Phase of the AI Boom – Business Insider — businessinsider.com
- Substack’s new tool tells you who’s been writing their newsletters with AI | TechCrunch — techcrunch.com
- From Substack to YouTube, here are the social platforms cracking down on AI slop | Business Insider Africa — africa.businessinsider.com
