Substack partners with Pangram to add AI detection tool to its platform

My take: Substack added an AI detection tool today in partnership with Pangram, available now on web and iOS. Any reader can scan text of more than 100 words to estimate how much was AI-generated; creators will also have a space to explicitly disclose how they used AI in their content. Substack's CEO called the practice of using AI to fake human connection "Claudefishing."

For those of us who create content, this is a clear signal: transparency about AI use is no longer optional. The moment a platform itself builds tools for readers to verify whether what they're reading is human, the question of authenticity stops being philosophical and becomes practical.

AI can make you a more efficient creator, but if your audience cannot tell whether the content is yours or a model's, you lose something that is not easy to recover: trust. Do you already have a clear policy on how you use AI in your work, and how you communicate that to your audience?