AI Writing, Human Experience and the New Content Divide

Why this matters now

The conversation around AI writing is too often framed as a moral panic: human writers good, AI writers bad. That is not how I see it. I have spent years in ecommerce, SEO, Authority Building and automation, and I do not believe businesses win by pretending tools do not exist. They win by understanding what tools are good at, what they are terrible at, and where human responsibility must stay in the loop.

The article that triggered this response, published by Search Engine Journal, made an important observation: three different stories from different worlds all point to the same tension. In education, students can now produce fluent text without going through the struggle of forming their own ideas. On the web, AI-generated articles appear to have reached a massive share of newly published content. In the creative economy, freelancers face pressure from budget cuts and uncertainty about AI.

For me, this is not only a content marketing story. It is an SEO operations story. Search engines, AI answer systems and users are all becoming less impressed by text that merely looks complete. They need content that answers a real question in a specific context. They need evidence. They need examples. They need clear authorship, useful structure and a reason to trust the source.

This is especially important for SMEs. Small businesses do not have the luxury of publishing endless articles that sound clever but do not move customers. A local clinic, florist, ecommerce store, hotel, parking service or B2B provider needs content that helps a real buyer make a decision. A page about “best pediatric clinic in Bucharest” should not sound like a generic directory. It should help a parent compare care options, understand urgency, evaluate reviews, see location and parking issues, and know what to do next.

AI can help build that kind of content faster. But only if the business gives it real substance. Without real substance, AI simply produces more of the same: smooth sentences, weak differentiation and pages that feel written for no one in particular.

The three signals behind the new content divide

The first signal comes from education. In The Guardian, Micah Nathan, a novelist and MIT lecturer, wrote about confronting students who had used AI for creative writing. His point was not only about cheating. It was about the value of the writing process itself. Writing forces a person to discover what they actually think. If the machine finishes that struggle too early, the final text may look acceptable, but the thinking behind it can be absent.

That matters for SEO because the same absence is visible in commercial content. You can feel when a page has no field experience behind it. It defines a topic, lists predictable tips and concludes with generic advice. It might not be technically false, but it does not help enough. It does not show that someone has solved the problem before.

The second signal comes from the web itself. Graphite’s analysis of Common Crawl article URLs suggests that primarily AI-generated online articles rose sharply after ChatGPT launched, then stabilized around roughly half of new articles. The important lesson is not that “AI has taken over everything.” It has not. The important lesson is that the web now contains enough AI-shaped content that sameness becomes a strategic risk.

If half the market can generate text at roughly the same level of fluency, fluency stops being a moat. The new moat is specificity: first-hand examples, original interpretation, operational insight, proprietary data, expert review, direct experience and the ability to turn content into better website structure.

The third signal comes from the people doing creative work. Research reported by The Accountancy Partnership and covered in industry media shows that creative freelancers are dealing with budget pressure, stress and uncertainty around AI. That economic pressure can push the market toward cheaper, faster and less differentiated content.

This creates a dangerous cycle. Clients cut budgets. Writers are asked to produce more for less. AI is used to compensate for lower fees and tighter deadlines. The output becomes more generic. Generic output performs worse. Clients then lose even more trust in content. The solution is not to abandon AI. The solution is to stop using AI as a substitute for judgment.

Google’s quality test is not anti-AI. It is anti-empty content.

Google’s own documentation does not say that AI-assisted content is automatically bad. What Google repeatedly emphasizes is helpful, reliable, people-first content. Its self-assessment questions ask whether the content provides original information, analysis, value beyond the obvious, evidence of expertise and a satisfying answer for the reader.

This is the right way to think about AI writing. The question is not “Was AI involved?” The question is “Does the final page deserve to exist?”

A page deserves to exist when it gives the reader something they could not easily get from ten other pages. That might be a practical framework, a real case, a local market example, a tested workflow, a comparison table, a warning from experience or a clear next step. It might be a better explanation of a technical issue. It might be a checklist that actually maps to implementation.

Google also talks about E-E-A-T: experience, expertise, authoritativeness and trust. I do not treat E-E-A-T as a magic checklist. I treat it as a discipline. If the article is about SEO for florists, show that you understand florists. If the article is about AI visibility for ecommerce, show that you understand product feeds, category pages, stock, reviews, pricing, shipping and buying intent. If the article is about backlinks, explain relevance, risk, placement context and authority signals rather than simply saying “links are important.”

AI can help assemble the first draft of these ideas. But AI cannot invent your actual operational experience. If it does, it is likely to hallucinate or generalize. That is why a serious AI-assisted content workflow must include research, fact-checking, human editing and post-publication execution.

What SMEs should learn from this

Most small and medium businesses should not build a content strategy around volume alone. That was already risky before generative AI. In the AI Search era, it is even weaker. If everyone can publish more, publishing more is not the advantage.

The advantage is knowing what should be published, why it matters, how it connects to the rest of the website and what changes need to happen after publication.

For example, if a parking business near an airport publishes a guide about “where to park near the airport,” the article should not only define parking options. It should compare distance, shuttle reliability, security, booking flow, cancellation rules, luggage concerns, late-night arrival and price transparency. It should link to service pages, answer objections, support local search visibility and create a page that both a human and an AI answer engine can understand.

If a medical clinic writes about a symptom or treatment, the stakes are higher. The content must be reviewed carefully, avoid overclaiming, cite appropriate sources and make clear when a patient should contact a professional. A thin AI article in a YMYL category is not only weak SEO. It can become a trust problem.

If an ecommerce store writes about product comparisons, the useful article should mention real criteria: sizing, compatibility, shipping, returns, use cases, materials, warranty, customer questions and what happens after purchase. Generic “best product” content is easy to generate and easy to ignore.

That is the content divide I see forming. On one side: fast, generic, superficially polished content. On the other: useful, specific, evidence-backed content connected to website execution. The second side is harder. That is why it will matter.

My AYSA view: AI should reduce busywork, not remove responsibility

AYSA was not created because I believe business owners should outsource their thinking to a machine. Quite the opposite. AYSA exists because business owners are drowning in SEO work that rarely turns into execution: audits, dashboards, keyword exports, technical issues, meta rewrites, internal links, content plans and recommendations nobody has time to implement.

The right role for AI is not to replace expertise. The right role is to make expertise operational.

In practical terms, that means the agent can monitor a website, detect opportunities, prepare changes, explain why they matter, ask for approval and then execute accepted actions inside the website workflow. The business owner or marketer does not need to manually copy every recommendation from a report into WordPress. But they should still understand the important decisions.

For content, that means AYSA should help answer questions like:

• Which pages get impressions but fail to satisfy the search intent?
• Which topics are missing from the site’s topical map?
• Which content needs human experience, local examples or buyer proof?
• Which articles should link to which service or product pages?
• Which FAQ, schema or internal linking improvements should be prepared?
• Which content is too generic to compete in classic search or AI answers?

This is where AI becomes useful for SMEs. Not as a machine that floods the site with articles, but as an execution layer that keeps moving the right work forward.

A practical workflow for AI-assisted content that can actually rank

Here is the workflow I recommend for SMEs that want to use AI without creating a pile of generic pages.

Start with a real business question

Do not begin with “write me an article about X.” Begin with the customer problem. What is the buyer trying to decide? What fear, confusion or comparison is behind the query? What does the business know that a generic model will not know?

Use search data, not only prompts

Search Console, Analytics, Business Profile data, rank tracking and customer questions should shape the brief. AI is much more useful when it is grounded in actual demand and website context.

Add human proof before publishing

Every serious article should include something that came from experience: a case, a pattern seen in client work, a practical limitation, a local market detail, a screenshot, a quote, a workflow or a decision framework. Without that, the page risks sounding like every other page.

Make the content structured for humans and machines

Use clear headings, concise definitions, comparison tables, summaries, examples and internal links. This helps readers, Google and AI retrieval systems understand the content.

Connect the article to execution

Publishing is not the end. The article should link to related pages, support product or service pages, trigger on-page improvements, update internal linking and become part of a cluster. This is where many content strategies fail: they publish and move on.

Monitor and improve

After publishing, watch impressions, clicks, ranking movement, AI visibility, engagement and conversion behavior. If the article gets impressions but poor clicks, improve the title and answer alignment. If it ranks but does not convert, improve proof and next steps. If AI systems cite competitors instead, look at entity clarity, authority and source quality.

My conclusion

The future of content is not human-only and it is not AI-only. It is accountable content. Content that can explain why it exists, who it helps, what experience supports it and what action should happen next.

AI will make average writing cheaper. That is already happening. But it will also make real expertise more visible, because the contrast between generic language and lived experience will become sharper.

For SMEs, the opportunity is not to publish more noise. The opportunity is to use AI to remove the operational bottlenecks that stopped good SEO work from being implemented. Research faster. Prepare better. Approve clearly. Execute consistently. Then improve based on real performance.

That is the side of the content divide I want AYSA to stand on.

Sources and further reading

• Search Engine Journal: 3 Unrelated Stories About AI & Writing Tell The Same Story
• The Guardian: Micah Nathan on AI and writing students
• Graphite: AI now writes as many online articles as humans do
• The Accountancy Partnership: The Industry Frustration Report
• Google Search Central: Creating helpful, reliable, people-first content
• Google Search Central: Guidance on using generative AI content
• AYSA: Scaling AI content without penalty
• AYSA: Quality content in the AI Search era