The Uncertain Landscape of AI Content on Medium

In the wake of numerous complaints from writers regarding AI-generated material, Medium has finally responded to calls for clarity. However, the news is not encouraging for those who consider Medium a refuge for authentic writing.

Medium's update reveals a significant issue: they confess to struggling with the identification of AI-generated content, leading them to believe that establishing a strict policy would be futile. They question the value of rules that cannot be enforced, stating, "What's the point of a rule you can't enforce?"

Their message is clear: writers should cease their grievances as they have implemented a system to suggest quality articles. This system relies on curators reviewing a mere 0.5% of published stories to unearth noteworthy pieces. If Medium overlooks 99.5% of the submissions, why should we give them our attention?

Currently, Medium indicates that while humans can detect content generated by tools like ChatGPT, automated systems struggle with this task. Curators feel confident in filtering out AI-generated pieces from the small percentage they review.

The platform's recent focus on publications and editors in the Boost nomination pilot highlights that we, as writers, are the first line of defense. Curators serve as the secondary line, but there’s a troubling twist: Medium has chosen not to enforce stringent rules against AI-generated content because they find it increasingly difficult to differentiate between human and AI narratives. They admit, "There's no real way for us to prove anything," and express growing uncertainty about their ability to identify AI content.

This scenario is concerning. It doesn’t take a visionary to predict that if we can’t distinguish between human and AI-generated writing, the unregulated 99.5% will overrun the curated 0.5%. Consequently, Medium could become a platform solely populated by high-quality AI stories.

This potential outcome is not just a possibility; it’s likely to happen. Whether intentional or not, Medium's current approach risks its relevance as a haven for human authors, especially as AI technologies like ChatGPT continue to advance.

My perspective is not as pessimistic as Medium's. Having managed compliance in a global corporation for years, I understand the magnitude of adhering to countless laws and regulations. Despite knowing that we wouldn’t catch every infraction, we never considered giving up and claiming, "It’s hard, so let’s not even try."

Writers, while Medium has listened, they seem oblivious to the reality we face if we allow this situation to persist. Their path could lead us on a one-way trip on a sinking ship. We must respectfully decline this trajectory and urge them to revisit their approach.

If Medium fails to provide a more pragmatic response, it’s reasonable for any writer with a sense of self-preservation to seek alternative options. Conversely, you might choose to join the ranks of those creating what I predict will soon be the majority of AI-generated content. It appears there will still be a place for you on Medium.

Take care.

This is an opportunity for me to mention my future newsletter on Substack. Though I don't have one yet, feel free to subscribe for updates once I launch.

• To be fair, Medium does not view their approach as ignoring the issue. Terrie’s response to my concerns can be found here: Comments section on Terrie Schweitzer's story.