I'm just going to be blunt about this. I'm seriously worried about the state of the internet. It's always been under assault from a variety of malefactors, bot farms, and obtuse algorithms, but right now, with AI being as dominant as it is, the very essence of quality journalism is under threat.

Not just tech journalism either, but the wider arena in general. It doesn't matter what is covered—tech, politics, sciences, health, fitness, or any of the other multitude of hobbies out there—it's all under threat. All thanks to a perfect storm of pseudo-monopolized search engine dominance and AI overviews.

Good, honest, journalism is and always has been the fourth cornerstone of all solid democracies. At its core, it's designed to hold power to account. To shine a light on the comings and goings of those in charge, those who would do wrong, those who would manipulate and abuse the general public, either overtly or in the shadows. 

Any student of journalism should be able to tell you that our work needs to satisfy three separate criteria. It needs to inform, it needs to educate, and it needs to entertain. To what degree each of those elements ranks varies from rag to rag, but inherently that's what we do. And tech journalism is no different in that regard. Reviews, features, builds, news, investigative reporting—all of it is predicated around this same notion.

The Homogenization of the Internet

In recent years, most tech sites have become remarkably similar. Not just in layout, but also in how they operate. Many do in fact sing from the same hymn sheet (not all), as they're often owned by the same massive publishing houses, who push their own content strategies down the pipeline to improve revenue generation and such. I'm not saying this is a bad thing, as it keeps people employed, and we still have fantastic journalists working across the board, but it's that homogenization in strategy that's made web journalism incredibly vulnerable.

The strat, at its core, is a fairly simple one. Reach the largest number of people possible via search, and then earn a decent revenue through affiliate links, either via buying guides ("the best power supplies of 2025" etc), or embedded links in content in a similar vein. Sponsored articles similarly follow suit as well, as manufacturers pay top dollar to secure dedicated articles on those same sites, enticed by insane levels of site traffic and more.

It's been an effective strategy for well over a decade now, and as it's grown, publishing houses have time and again purchased more and more of those smaller sites, bringing them into new templates, streamlining teams, and effectively pulling them onto that same strategy.

As a result, the SEO marketing side of the industry has exploded. Experts are everywhere, getting paid top dollar to ensure your site rankings improve on Google, targeting specific keywords, and ensuring traffic increases as a result. If the search volumes are right (the number of people searching for a specific term on Google), and you go up a rank utilizing that same term, you can see traffic double, which leads to more money for the site, the publisher, and well, those shareholders.

But more money means more journalists and better quality journalism.

AI Overview

The problem is that sites have become incredibly vulnerable to any major changes in the SEO algorithm that Google implements (and Google has 89.54% of the market share as of June 2025). 

SEO strategy, in its very nature, is designed to second-guess what markers Google uses to rank a webpage for a term and then take advantage of those to bolster said ranking. Typically through experimentation. If you've ever seen headlines like "I've tested every X product; this is what I'd buy" on repeat, this is why. Might look awful, but it's a game editors are playing to compete with each other on search that Google has secretly said works.

If Google suddenly decides it doesn't like a particular strategy sites and SEO experts are using, it often changes and readjusts the algorithm appropriately, which in turn can demolish and even destroy a website's traffic overnight, unless they rapidly adapt. That then eliminates revenue, page views, sponsorship, and more.

What's worse, Google can even just outright ban a site from being listed on its search engine through DMCA complaints too, if it so chooses. In fact we saw this most recently with HouseFresh, which Linus Tech Tips covered here for their review of a PuroAir product.

Then there's Gemini. On the surface a wholly promising tool implemented by Google to aptly sum up quick queries or complex questions. At the top of a search term, way above rank 1, Gemini will often give you an AI Overview, summing up whatever your query may be. Broken down into bullet points as LLMs often do.

Benign on the surface, it sources the information from websites, takes extracts of that copy, works it into a sentence or bullet point that makes sense, and then ties a little link to the end of each bullet point for the querier to click if they want more information.

The thing is, they often don't, that quick summary is (not always but) usually more than enough to answer the question at hand, and as a result, search traffic has absolutely fallen off a cliff. Some of the biggest sites on the web at this point are seeing massive drop-offs in traffic, affiliate sales are down, and as a result budgets are being cut dramatically to compensate for the sudden loss.

This doesn't just apply to tech either, but also things like health queries, science news, cooking recipes, political history, the works. Even sites that don't rely on affiliate links but do benefit from search traffic and SEO are getting slammed, and as a result, cutbacks inevitably have to be made.

The Snowball Effect

What you're left with then are far fewer journalists and specialists in the field. Experts who've been reviewing hardware or writing scientific papers can no longer afford to do it, and sites can't afford to employ them.

But it's a catch-22, because AI LLMs aren't capable of reviewing hardware, they can't perform experiments, they're not sat in the senate, in a lab, in a court room, or out in a war-torn country. As I see it, there's only one way this is going to go; inevitably, expertise, experience, authority, and trust will evaporate from these platforms. At varying levels and at different speeds, dependent on how tied into search and affiliates each brand is, but it is going to happen. Certainly as sites will no longer be compensated for that information.

When it does, AI is invariably going to become "dumber" as a result. It won't have that wealth of up-to-date knowledge to draw on, and over time the web will inherently become a quieter, darker place, with less accurate scrutinized information available, almost regressing back to its infancy.

Quality journalism and in-depth content, tailored to SEO, will be a thing of the past. Instead, parasite sites, AI-generated garbage, churnalism speculating on topics, and marketing material alone could potentially take over as the primary sources of knowledge that an LLM draws from to create its AI Overviews, in turn accelerating the cycle.

From there, you'd assume (or hope) that people, recognizing that lowering in quality in AI responses, and lack of online content, would similarly leave the platform as well, as their queries and questions are no longer answered competently enough, by AI or otherwise. Pushing us ever further towards a dead internet.

Is there any hope?

There's an irony, I get it. I know I'm an ex-print journalist from a magazine that's just been closed. I'm not going to pretend that they're the answer (arguably the fact that Maximum PC survived as long as it did was a miracle), yet the way I see it, this shift is remarkably similar to the general journalistic upheavals we've had in the past.

From print to radio, from radio to TV, and TV to web, each form of journalism has been altered, often radically, by new technologies. AI in its current capacity feels very similar to that. But, paradoxically, it seems that instead of these massive publishing houses banking on one strategy, on SEO and affiliates, that what we may end up with is almost a reworking of the internet, away from Google's dominance, and instead back to a URL-dependent world, where we visit websites and journalists and we support good-quality content through choice, rather than through subversion and affiliates or ad revenue, and although the transition will inevitably be horrific, the scale-backs dire, and the effect on the livelihoods of millions disastrous, perhaps there's light at the end of the tunnel. At least, we can only hope.

The link has been copied!