AI-generated content is increasingly appearing on platforms like Medium, a 12-year-old publishing site that has recently turned a profit. Medium CEO Tony Stubblebine promotes the platform as a "home for human writing, " but reports indicate that nearly half of recent posts may be AI-created. An analysis by Pangram Labs showed that over 47% of 274, 466 sampled Medium articles were likely AI-generated, significantly more than other internet platforms, where the average is about 7%. Tags such as "NFT" and "web3" featured heavily in AI-generated content, with nearly 78% of articles labeled with "NFT" flagged as likely AI. A comparison by Originality AI found that AI content on Medium rose significantly from 3. 4% in 2018 to over 40% in recent samples.

Despite this data, Stubblebine disputes the notion that Medium struggles with AI content, arguing that the impact of AI-generated posts is minimal since many receive zero engagement. Medium has taken measures to reduce AI-generated content, updating its policies to not allow AI writing in its paid Partner program and limiting the visibility of undisclosed AI content. Stubblebine asserts that existing moderation systems help filter out low-quality posts, as much AI writing also tends to be subpar. While some writers appreciate Medium's approach to quality control, others argue that AI-generated pieces still saturate the platform. The challenge for Medium is not unique; it reflects a broader issue on the internet, where the rise of AI complicates the landscape for original human-created content. The ongoing situation hints at a future where the web may increasingly feature AI-generated material over authentic human contributions.