Substack Introduces AI Detection Tool and Optional AI Author Notes

Substack has introduced new AI transparency features through a partnership with AI detection company Pangram, giving readers the ability to estimate how much of a post was written by AI while also allowing creators to voluntarily disclose how they used AI in their work. The new tools are available beginning today on the web and iOS, with Android support planned for a future release.

The AI detection feature lets users scan posts, comments, replies, and other content published from today onward. According to Substack, the tool works on text longer than 100 words, while the company also said the feature will be available for posts, notes, replies, and comments exceeding 100 characters within its app.

Alongside the detection tool, Substack has added an optional AI Author's note that gives writers a dedicated space to explain whether AI played a role in creating their content. The company said the goal is to encourage creators to share their process rather than discourage AI-assisted writing.

Substack described the new capabilities as a way to help readers understand what they are consuming, saying its guiding principle is that "people should know what they're getting."

The company also acknowledged that AI detection has technical limitations. It noted that Pangram cannot reliably identify every instance of AI-generated writing, reflecting broader challenges with current AI detection technology. Publishers who believe a scan incorrectly identified their work can run Pangram on drafts before publication and report or remove scans they consider inaccurate.

Substack's announcement also took aim at the broader online landscape, criticizing the rise of AI-generated content on other platforms and referring to attempts to simulate human connection with AI-generated material as "Claudefishing."

Chief executive Chris Best framed the feature as supporting, rather than replacing, human creativity. "This is good use of AI," Best said.

The rollout follows a broader industry shift toward labeling AI-generated content across online platforms. Social media services have added labels for AI-generated images and videos, while music streaming platforms have introduced disclosure policies and, in some cases, measures targeting AI-generated music.

Substack said its AI detection tools are intended to improve transparency rather than enforce restrictions on AI-assisted writing. By combining automated detection with optional creator disclosures, the company is positioning the feature as a way to help readers better understand how content is produced while leaving publishing decisions in the hands of writers.

This analysis is based on reporting from engadget.

Image courtesy of Substack.

This article was generated with AI assistance and reviewed for accuracy and quality.