- Substack is launching a feature with Pangram to detect AI-generated text, aiming to maintain human authenticity.
- The problem is not AI use per se, but the mismatch between reader expectations and reality, especially when readers unknowingly engage with AI content.
- Substack is introducing tools for transparency, including a detection scan and a 'How I make this' statement for creators.
- The feature is available on web and iOS, with Android coming soon; future plans include customizable preferences for AI content.
- The article highlights a debate: detection tools can help transparency but may become a purity test, as AI-assisted writing can still involve human creativity.