Substack is betting AI labels will protect trust before AI slop erodes it

Substack is betting AI labels will protect trust before AI slop erodes it
Substack is making a pointed bet about the future of online writing: readers will keep paying for newsletters only if they can tell whether a human actually wrote them. Its new partnership with Pangram turns that bet into a product feature, letting users scan posts, comments, and replies for an estimate of AI-generated text.

The move is less an anti-AI crackdown than an attempt to draw a clearer line around authorship. Across coverage of the launch, the central argument is consistent: Substack wants to preserve trust as AI-assisted publishing becomes harder to spot. TechCrunch framed it as a transparency play that could hurt in the short term by exposing how much content on the platform may not be fully human-written, but help in the long term by keeping out what it called “AI slop.” Axios described the strategy more bluntly, saying Substack is betting subscribers and advertisers still “value authenticity” and will pay for work “verifiably created by humans.”

Substack’s own position, as relayed across outlets, is that the real problem is not AI use itself but confusion about what readers are consuming. CEO Chris Best argued that “when content made by no one takes over parts of the internet that are supposed to be human, it pollutes the commons and makes it hard to discover and hear human voices.” The Verge highlighted his sharper formulation: “The core problem is not people using AI, or the quality of its output,” but “a mismatch between a reader’s expectation and reality.”

That distinction matters because Substack is pairing detection with disclosure. In addition to scans, writers will be able to add a “How I make this” note explaining their process, and creators can challenge scans they believe are wrong. Best also stressed the limits of the technology, noting Pangram can detect signs of AI use but not “whether great human care went into creating it.”

In other words, Substack is not trying to ban AI from its platform. It is trying to make authorship legible again — before readers decide too much of the internet sounds like nobody at all.

Continue reading https://foxvector.com/stories/019f9fce-e1ce-11fe-70b1-36adf7fcd548