The AI content reckoning isn’t coming from a policy update. It’s coming from readers who are tired of reading nothing.
There is a version of this story that makes Google the villain. The algorithm changed, the goalposts moved, companies got punished for doing what everyone told them to do. I have seen that story make the rounds this month and I want to push back on it, because I think it misreads what is actually happening.
Google did not announce that human writers are back. What they actually said, and have consistently said, is that they do not care whether content is AI-generated or human-authored. What they care about is whether it is original, trustworthy, and demonstrates first-hand expertise. That standard is called E-E-A-T: Experience, Expertise, Authoritativeness, Trustworthiness. It is not new. It has been the framework for years.
What changed is that a flood of content arrived that fails every one of those four tests simultaneously, and the gap became impossible to ignore.
The problem was never AI. The problem was what people did with it.
I have watched a lot of companies approach AI content the same way. They open ChatGPT, type something like “write me a blog post about [our product category],” copy the first output, and post it without a second read. Some of them do this at scale. Hundreds of posts. Entire content calendars built on that loop.
The output is technically coherent. It has headings. It has paragraphs. It answers the question it was asked in the most predictable possible way. It sounds like no one, because no one was actually there.
Google has not penalised that content yet in the way it eventually will. But readers have already started leaving. The signal is quieter than a ranking drop. It shows up in time on page, in bounce rate, in the absence of replies and shares and the kind of secondary traffic that comes from someone forwarding something to a colleague because it said something real. That traffic is disappearing from AI-slop sites whether or not the rankings have moved yet.
The rankings will follow. They always do.
What Google actually launched at I/O this year.
The announcement that got misread as a “human writing comeback” was actually something more interesting. Google introduced a tool that acts as a thought partner, something that helps a writer turn a rambling set of ideas into a structured first draft. Not a replacement. A collaborator. The distinction matters enormously.
That framing is not accidental. Google is signalling what good AI-assisted writing looks like: a human being with real experience and genuine perspective, working with AI to get it out faster and more clearly. The experience still has to be there. The perspective still has to be there. The AI is the mechanism, not the source.
The exoskeleton, not the autopilot.
I write this way and I want to be direct about it, because I think the transparency is worth something.
The articles on this site are written through a process where I answer questions, usually over breakfast before the meetings start, and the draft comes from those answers. The questions I get asked are sometimes ones I would never have thought to answer on my own, and that friction is part of what makes the output useful. What goes into the piece is twenty-plus years of working inside growth teams at companies like Copper, Certn, and Embroker, plus a decade of advisory work through VonClaro. The AI doesn’t have that. I do. The AI helps me lift it.
The best analogy I have found for this is an exoskeleton. The 500-kilogram block I couldn’t realistically move before becomes manageable because the structure is there to help. I am more productive. The output is better than I could produce alone in the same window of time. But the thing doing the lifting is still me.
That is a fundamentally different product than a company pasting the first output from a prompt and calling it content strategy.
What happens to the raw-AI sites over the next twelve months.
The companies that went heavy on unedited AI content are not in trouble yet. The rankings haven’t moved dramatically for most of them. But the editorial debt is accumulating and the window to fix it without serious pain is closing.
My read: the sites that run genuine editorial on their existing AI content, bringing in real writers, grounding the claims in actual expertise, adding the perspective that was missing, will stabilise. The ones that don’t will see rankings drop in a way that is difficult to reverse, because the content library they built provides no foundation to recover from. You cannot compound on generic.
The companies that figure this out first and move to a real hybrid model, human expertise with AI-assisted production and genuine editorial standards, will not just hold their position. They will gain ground on competitors who are still debating whether to act.
E-E-A-T was never the hard part. The hard part was always having something worth saying.
This article was written in the way it describes.