Some writers and editors on Medium commend the platform’s approach to integrating AI. Eric Pierce, the founder of Medium’s largest pop culture publication, Fanfare, notes that he rarely encounters AI-generated submissions. He believes that Medium’s human curators, through its boost program, play a crucial role in spotlighting the best human-authored content. “I haven’t read a piece on Medium in recent months that even hints at being AI-generated,” Pierce states. “Medium increasingly feels like a stable refuge amid a chaotic internet landscape.”
Conversely, a segment of writers and editors reports encountering substantial AI-generated content on Medium. Marcus Musick, a content marketing writer and editor of multiple publications, expressed his concerns in a post about an article he suspects was AI-generated, which gained significant attention. Reality Defender’s analysis suggested the article was “99 percent likely manipulated,” accruing over 13,500 “claps.”
Musick notes that as an editor, he believes he frequently encounters AI-generated content, rejecting about 80 percent of monthly submissions suspected of being AI-produced. Instead of relying on AI detectors, which he considers ineffective, he relies on his own judgment.
The prevalence of AI-generated content on Medium underscores the ongoing moderation challenge: distinguishing quality content from subpar submissions—a longstanding issue across the internet, only exacerbated by the surge in AI use. While click farms have posed problems for some time, AI facilitates a new means for SEO-focused entrepreneurs to quickly revive dormant media platforms with low-quality AI content. A subgenre exists among YouTube entrepreneurs who publish tutorials encouraging the creation of AI-generated content across various platforms, including Medium.
Jonathan Bailey, a plagiarism consultant, observes that Medium mirrors the broader internet crisis, where AI content proliferates rapidly. He suggests that tools like spam filters and human moderators are probably the best defenses against this influx of content.
Stubblebine argues that the presence of low-quality content on a platform is less significant than the platform’s ability to amplify quality writing while restricting the dissemination of inferior material. His approach to moderation may prove strategic in this context.
This situation hints at a possible future aligning with the “Dead Internet” theory. This theory, rooted in online conspiratorial thinking, posits that much of the internet lacks genuine user engagement, instead being dominated by AI-generated content and bots. As generative AI becomes more prevalent, platforms that do not actively combat bots may foster an environment where human-created work becomes increasingly scarce amidst the overwhelming presence of AI content.