I found three different articles about writing and AI in the same week, each from a completely different angle, and all explaining the same thing.
A novelist turned MIT writing teacher confronts students who have outsourced their essays to AI. A new study by Graphite shows AI-generated articles now make up nearly half of all new content on the web and are prevalent there. And new data from The Accountancy Partnership shows that half of independent creatives say increased stress is affecting their work, as client budgets for creative human resources shrink.
One data point is true. The second is mutual. The third is habit.
When read together, these articles form an argument that every SEO professional, content marketer, and creative freelancer should take seriously, acknowledging the content divide that is occurring and asking, “Which side are you on?”
First Story: What Happens When Students Strike Out
On May 10, Micah Nathan, a novelist and MIT lecturer in fiction and nonfiction writing, published a piece on. The guard about confronting his creative writing students with their use of AI. The confessional session that followed, he wrote, was one of the most productive teaching sessions of his eight years at MIT.
His key insight was not about academic honesty. It was about what it means to write. “Writing is not just about producing sentences,” he told his students. “Perseverance training in the form of continuous attention. It is a way of learning what a person thinks by trying to say it. LLM can reproduce the appearance of that work, but it cannot replace it, because the value is not only in the thing produced but in changing what happens during its execution.”
He described AI’s prose as “flawless, commonplace, strangely empty,” borrowing Tennyson’s description of a beautiful but empty face, producing what he called “a simulacra of thought, produced by pattern recognition learned from millions of human words, based on some knowledge not made by a single person.”
Smart students, he said, sense that you have nothing even if they can’t explain it.
For SEO experts, this is not a concern for the documentation. It’s an accurate description of the content quality problem that Google’s helpful content programs have been trying to solve since 2022. The sign that Google is hunting for is exactly what Nathan refers to as something that AI can produce – evidence of a mind working hard to deal with a particular problem arising from a particular experience. Pattern recognition learns from what people write. It cannot repeat why they wrote it.
→ Read More: Why Great Content Is No Longer Enough and What Beats You in Search AI
Story Two: The Dreaded Takeover Hasn’t Happened – Yet
On May 15, Megan Morrone reported Axios in new data from digital advertising agency Graphite, which analyzed 55,400 online articles and lists published between January 2020 and March 2026, using each of three AI discovery tools. The findings were more subtle than most AI content which was about the share of AI-generated content primarily, which held around 50% for over a year and appears to be plateauing.
The feared takeover has yet to happen. AI content briefly surpassed human-written content in late 2024, but the two have remained on par since then.
An important caveat Morrone included is that most articles are no longer written by humans or AI alone. One might use AI to express, write, rewrite, or edit, making the line really blurry. Dan Klein, UC Berkeley professor and CTO’s AI model, has flagged the danger of a feedback loop. If models are trained mostly on AI-generated content, the internet can become a machine that produces low-quality content that trains models that produce very low-quality content.
For SEO experts, the plateau is reassuring and warning in equal measure. The volume shock is extreme. But the problem of quality dilution is real and growing, and it creates the same opportunity that Nathan identified on the other side. On the web where almost half of the content is generated by AI, the content that contains real human knowledge and expertise is more diverse, not less.
→ Read More: AI Platform Founder Explains Why We Need to Focus on Human Behavior, Not LLMs
Issue Three: The People Who Produce This Content Are Under Heavy Pressure
On May 13, Emma Hull e Accounting Partnership emailed me directly data from a new report for creatives across the PR, marketing, performing arts, graphic design, photography, and related industries. Half of independent creatives (50.7%) say rising stress levels are interfering with their work. Half (50.2%) say customer budget cuts are the biggest challenge they face in 2025. More than two in five (43.3%) believe that AI will have a negative impact on their industry. About half work unpaid hours each week.
Lee Murphy, Managing Director at The Accountancy Partnership, put it bluntly: “Creative work is often heavily associated with marketing budgets and discretionary spending.
The paradox involved in these three numbers is worth pondering. Clients reduce the budget of human creative work at the same time AI generates almost half of the content on the web, while a professor at MIT writes some cognitive costs that remove the writing process without anyone doing it, whether you are a student or a professional.
Independents under the most pressure are the most tempted to use AI to produce more content quickly to compensate for lower prices. The content they produce thus becomes 50% inseparable from the output of the machine. And content that cannot be separated from machine output is exactly what Graphite data and Google’s quality systems train users and algorithms to reduce.
→ Read More: Over-Reliance on AI Holds Businesses Back
What the Pattern Really Means
These three stories, read together, describe the market in a bifurcation process. On the one hand there is always high-volume, low-variety content that is produced quickly, that is low-value, and increasingly difficult to distinguish from AI output, regardless of who produced it. On the other hand there is content that contains the specific expertise, specific knowledge, and editorial judgment that Nathan’s students were trying to avoid. Content that takes longer, costs more, and is increasingly the only type that gains meaningful search visibility and reader trust.
This is not a new argument in SEO. What is new is the powerful clarity that three independent sources from three completely different fields – literature education, web content analysis, and independent labor economics – all point to the same conclusion in the same week.
Shelley Walsh made the point in her recent Search Engine Journal piece about AI content ranking that the distinction between commodity and non-property is where the real strategic question resides. The above three stories are proof that the divide is already there, already measurable, and already affecting people’s lives.
Writers who understand this, and produce accordingly, are the ones who will still have work to do when the budget cycles turn again.
Additional resources:
Featured image: SvetaZi/Shutterstock
