Substack launches new tool to detect AI-written text

Substack, one of the world's largest platforms for independent journalism and newsletters, has introduced an unexpected update for its users. Readers will now be able to determine whether articles from the authors they follow were written by a human or generated by AI like ChatGPT. This step is aimed at maintaining content quality and increasing reader trust. This was reported by Techcrunch.com reports.

Substack implemented this feature by integrating the Pangram software. According to TechCrunch, the new tool scans posts, comments, and replies on the platform, providing an estimate of how much of the text was written by a human versus AI. This function applies to any text longer than 100 characters.

Transparency and author accountability

In the short term, this update might negatively impact Substack's business, as it will expose that many popular newsletters are actually written by bots. However, in the long run, this measure will help clean the platform of "AI slop" (low-quality AI-generated content) and build trust among readers regarding the information they consume.

Substack CEO Chris Best praised the use of this technology. According to him, software can handle all the technical work, but only a human can provide the ideas and emotions worth reading. "When I pitched Substack to writers, I used to say we do everything for you except the hard part. The hard part is the human thought and idea," Best emphasized.

Information, not a ban

It is worth noting that Substack does not intend to ban the use of AI. On the contrary, the company allows authors to add a special "AI Author’s note" to their texts. This enables creators to openly explain how they used technology in their workflow. Publishers can also check their drafts via Pangram before publishing to fix errors.

This trend is becoming widespread in the global tech world today. For instance, AI-generated images and videos are being labeled on social media, and music streaming services are placing restrictions on tracks created by neural networks. Substack is joining this movement, prioritizing authenticity in text-based content.

This update is also significant for users and authors in Uzbekistan. At a time when the use of AI tools is becoming popular among local content creators, such transparency tools will help distinguish high-quality and original Uzbek-language content.