Substack Launches AI Text Detection to Combat 'Claudefishing' and Protect Human Creators
Key Takeaways
- ▸Substack partnered with Pangram to launch AI detection across notes, posts, and comments, allowing readers to see transparency scores on content
- ▸The feature addresses 'Claudefishing'—misleading readers about AI-generated content—which Substack sees as a critical threat to platform trust and creator livelihoods
- ▸Substack is positioning itself as a human-first alternative to other social platforms, which reportedly contain as much as 40% AI-generated content
Summary
Substack announced the launch of AI text detection powered by Pangram, a feature now available on its platform to help readers identify whether content was generated or written by hand. The feature scans text longer than 100 words and provides a transparency metric showing how much of a post was likely created with AI assistance. The move addresses what Substack calls 'Claudefishing'—deceiving readers about AI-generated content—as an estimated 40% of content on some social media platforms is now AI-generated.
The feature is available on web and iOS today, with Android support coming soon. Substack is also introducing creator tools, including a 'How I make this' statement feature and the ability to run Pangram scans on drafts before publishing. The company framed the initiative as essential to maintaining trust on its platform, which positions itself as an economic engine for human creators and culture rather than a space dominated by soulless, AI-generated 'slop'.
While Pangram can detect AI-generated text with high accuracy according to independent research, Substack acknowledged the tool's limitations—it cannot determine whether great human care went into creating content or whether AI was used as a research source. The company indicated future possibilities including customizable AI content preferences in Reply Rules and additional transparency tools for creators.
- Creator tools include pre-publication scanning with Pangram and 'How I make this' statements to set reader expectations about content creation process
- The company acknowledged detection limitations and plans to expand features based on user feedback, including customizable content preferences
Editorial Opinion
Substack's move is a savvy differentiation strategy in an increasingly AI-saturated content landscape. By introducing transparency tools and positioning itself explicitly against 'Claudefishing,' the platform appeals to readers and creators who value human-generated culture and authentic communication. However, the initiative's success depends on whether readers actually engage with detection tools and whether transparency alone can counter the economic incentives driving AI-generated content proliferation. If Substack's approach gains traction, it could establish a new standard for how platforms balance AI capabilities with human authenticity.
