Parts of the SEO and GEO world love to demonize AI content. Andrew Holland, director of SEO at UK digital PR and SEO agency JBH, argues in Search Engine Land that much of that argument is self-serving: it protects an old business model, confuses production with quality and misses where the value of writing is moving. His piece reframes the debate about the value of AI content.
Many SEO agencies, he says, were really selling advertising. Content — produce it, rank it, drive traffic — fit the retainer model perfectly because it took time, writers were plentiful and agencies could charge well for production. Then AI made words incredibly cheap, creating an inconvenient problem for anyone whose margin depended on words being expensive.
When AI-fueled sites spiked and crashed, the industry said “we told you so.” But you can’t judge that from the final downward line. You need costs, returns and whether the economics worked. The real question is what the content was supposed to do and whether it did it.
Few readers finish an article and immediately buy software. As adman Howard Gossage put it, people read what interests them, and sometimes it’s an ad. The dull 2,000-word guide written because a keyword tool showed 1,300 monthly searches failed that test long before ChatGPT.
The assumption that human writing is automatically good is wrong. The $10 article with no research, experience or viewpoint predates generative AI. Holland describes a friend, a “Head of” at a SaaS company, who balked at a writer’s quote, built an AI content process that produced “good enough” drafts and had his team make them special. “I don’t use AI” isn’t a badge of quality; it can mean slower and more expensive. AI slop isn’t a writing problem. It’s an absence-of-ideas problem.
Search Engine Land reported that across 64 publisher sites, organic clicks fell 42% after AI Overviews expanded. Still, content is your company’s voice — what you believe, what your product does and doesn’t do, what you’ve learned and what you disagree with. People, search engines, AI systems and journalists can all encounter it. Say nothing, and the market has nothing to work with.
If AI can research and draft increasingly well, we’re paying writers for thinking: research direction, experience, taste, observation, argument, customer understanding, original information, editing and judgment — deciding what deserves to exist. Writers used to serve the keyword. Now they should move upstream:
Those beat “how many words should this be?” The scarce resource is no longer words but ideas worth communicating — and AI helps experts who struggle with prose get those ideas out.
Harry Potter and the Philosopher’s Stone started with a first print run of 500 copies. Years later, J.K. Rowling submitted The Cuckoo’s Calling as Robert Galbraith; publishers rejected it and one suggested a writing course. It sold around 1,500 hardbacks — until her identity was revealed and it hit No. 1. Same book, same sentences, different context. Quality isn’t an objective substance poured into a manuscript.
The question isn’t whether AI wrote it, but whether it was worth writing and reading. Content teams should shift time from production to the upstream work of securing original data, experience and perspective. Pair this with designing an AI content workflow backward from the finished piece and the slop antibodies platforms are building to see what will survive. Welcome to the idea economy.
Holland argues what matters is whether content is worth writing and reading, not who typed it. Low-quality human content always existed; AI slop is fundamentally a lack of ideas.
Upstream thinking: research direction, experience, observation, argument, customer understanding, original information, editing and judgment about what deserves to exist.
Yes. Content is the brand’s voice and raw material for people, search engines, AI and journalists — but it needs original perspective and information, not keyword filler.
To apply what you just read to your own site, start with a free audit of where things are now.
A strategist replies within 24 hours on business days.
Views without sales usually means you're making only one kind of content. A four-stage content funnel framework — with a skincare brand example, a pass/fail test for each stage, and a fill-in-the-blank planning sheet.
Organic clicks are down 42% since AI Overviews expanded. Traffic data from 30 major publishers shows top-of-funnel content collapsing in YMYL while growing in B2B tech — and points to three TOFU strategies that still work.
Search volume tells you what people search for, never why. Here is how to build content briefs from customer situations using the 7 W's, and how to prove the approach works.