eWEEK’s content and product recommendations are editorially independent. When you click on links to our partners, we may earn money. learn more.
Ten percent of the largest newsletters on the Substack publishing platform are likely to use some form of AI-generated content, and 7 percent of newsletters “heavily rely” on AI content, according to new analysis. It turned out that there was. Analysis was performed by GPTZero. GPTZero’s AI detection tool can determine the likelihood that a text was created by a generative AI. The results were published only in “WIRED.”
GPTZero studied the 100 most popular newsletters on the Substack platform and used an AI detection tool to scan 25-30 posts in each newsletter. GPTZero found that 10 of its newsletters were more likely to use AI-generated content created by ChatGPT and other platforms, and 7 of those publications relied heavily on it .
WIRED contacted all seven newsletter companies. The magazine said four of them admitted that artificial intelligence tools were part of their writing process. The remaining three did not respond to requests for comment. However, most of the publications that responded did not fully generate content for their Substack newsletters, but rather used AI tools for posting to social media, creating summaries, creating images, checking the grammar of posts, and other related tasks. You mentioned that you are using .
“I proudly use modern tools to increase productivity in my business,” says David Skilling, who runs Original Football Substack, which has 630,000 subscribers. “Although AI detection tools may detect the use of AI, there is a big difference between what is generated by AI and what is generated with AI assistance.”
Substack does not have an official policy regarding the use of AI on its platform, so there is nothing prohibiting writers from using generated AI to create content for Substack. Helen Tobin, Substack’s head of communications, declined to comment specifically on GPTZero’s results.
“We have several mechanisms in place to detect and mitigate fraudulent and organized spam activity, including copypasta, duplicate content, SEO spam, phishing, bot activity, and many more. may involve AI-generated content,” she told WIRED in an email. . “However, because there are many valid and constructive applications for AI-assisted content creation, we do not actively monitor or remove content based solely on its AI origins.”
To put the Substack AI news in perspective, another analysis by GPTZero found that approximately 1 in 20 Wikipedia articles uses content created by AI. This is about half the percentage of Substack’s results. However, additional analysis with two other AI detection tools found that nearly 40% of posts on the Medium publishing platform contained AI content. The difference is that while most of Medium’s posts get little engagement, Substack’s perpetrators represent some of the most popular newsletters on the platform.
