Substack does not have an official policy regulating the operate of artificial intelligence. One of the co-founders of Substack, Hamish McKenzie, described the boom of generative artificial intelligence as a radical change that writers will have to grapple with, regardless of their personal views on the technology: “Ultimately, it doesn’t matter whether you’re for or against this development. It’s happening,” he wrote in a Substack post last year.
Several Substack authors interviewed by WIRED emphasized that they used AI to refine their prose, rather than generate entire posts. David Skilling, CEO of the sports agency who runs the popular football newsletter Original Football (over 630,000 subscribers), told WIRED that he sees AI as a deputy editor. “I am proud to use modern productivity tools in my companies,” says Skilling. “AI detection tools can detect the use of AI, but there is a huge difference between AI generated and AI.”
Subham Panda, one of the authors of Spotlight by Xartup (over 668,000 subscribers), which covers startup news from around the world, said his team uses artificial intelligence as “an assistive medium that helps us curate high-quality content faster.” He stressed that the newsletter primarily uses artificial intelligence to create images and aggregate information, and that authors are responsible for the “details and summaries” contained in their posts.
Max Avery, author of the financial newsletter Strategic Wealth Briefing With Jake Claver (over 549,000 subscribers), says he uses AI-powered writing software like Hemingway Editor Plus to polish rough drafts. He says these tools help him “do more work creating content.”
Financial trader Josh Belanger says he similarly uses ChatGPT to streamline the writing process for his Belanger Trading newsletter (over 350,000 subscribers) and relies on the Claude chatbot to help him with editing. “I’ll write down my thoughts, my research, anything I want to include, and I’ll put it online,” he says. Belanger also creates custom GPTs (task-specific versions of ChatGPT) to help refine more technical text that includes specific jargon, which he believes reduces the number of hallucinations generated by the chatbot. “There are a lot of nuances to writing about finance or trading… The AI won’t know about it, so I have to prompt it,” he says.
Compared to some of its competitors, Substack appears to have relatively little AI-generated text. For example, two other AI detection companies recently found that nearly 40 percent of content on the blogging platform Medium was generated using AI tools. However, much of the questionable AI-generated content on Medium has seen low engagement and readership, while AI texts on Substack are published by powerful accounts.