Artificial intelligence tools are transforming the way people find information on the internet and this change is happening faster than publishers can adapt. When users question chatbots rather than using search engines, they are given answers, rather than links to follow and this is changing the economics of the internet – and in particular the economics of news providers – who spend money gathering information that they can no longer monetize.

Over the last few years, millions of users have shifted from search engines to AI chat tools for research, recommendations, and real-time answers. Tools like ChatGPT, Claude, and Perplexity are now directly answering questions that once sent readers to primary online sources that they felt they could trust. As users drift away from trusted news sources, they place growing faith in AI systems trained to mimic authority—tools that scrape the web for answers but offer no accountability.

About a year ago Google rolled out AI Overviews and then more recently AI Mode - features that answer questions directly on the google search page, often without crediting the original sources. For users, this feels seamless – they don’t have to constantly refine their search terms to find the information they were looking for and they don’t have to read ten different articles to work out an answer to their question - but for publishers, this has been disastrous.

A report by Enders Analysis, based on Sistrix data was released a few weeks ago showing that news visibility in search results has collapsed. The Mirror’s presence on Google is down 80% since 2019. The Mail has lost more than half. Even the Financial Times – a specialist publication with a loyal subscriber base, saw a 21% drop in traffic this spring. The report attributes this decline primarily to changes in Google’s systems, in particular - the rollout of AI Overviews and AI Mode, rather than publisher strategy.

This shift is structural. AI tools are intercepting the audience before they reach the source of the information they are seeking. Google’s AI overviews is doing this even for users who weren’t looking for an AI answer. Publishers can’t opt out from this without disappearing from search entirely – meaning that the economic model that sustained journalism—clicks, subscriptions, advertising—is being rapidly eroded by systems that extract value from news providers without returning any.

This chart from the Economist shows that news and media are by no means the most affected by this change. Science and education sites and reference sites (like Wikipedia) are doing even worse – with health information being the most impacted.

Search engines were – and still are – to a certain extent - the gateway to the internet – but that gate is rapidly closing. The shift in user behavior is already measurable. In April, Apple reported the first-ever decline in Safari search volume, a change they attributed to users turning to AI chat tools instead of traditional search engines. A TechRadar survey conducted in December found that 27% of U.S. users and 13% of UK users now use AI tools like ChatGPT instead of search engines to begin their information gathering. The reasons cited include speed, specificity, and ease of use.

Google referrals to news sites have dropped from around 65% in 2019 to 30% today, but this decline began long before AI Overviews.

A few years ago, people used to say that “nothing ever disappears from the internet” – usually as a warning against posting anything that you might later be ashamed of – but we have since learned that things can still disappear – even if they are never erased because they just become impossible to find.

Over the last few years, Google became less and less efficient at finding the information you were searching for – both because the web became saturated with SEO-optimized content, and because google started returning sponsored search results. Users found that they had to sift through clutter to find credible sources.

AI tools turned up on the scene right as internet search was faltering and according to The Atlantic - AI-chatbots have already replaced search for more than 25 percent of Americans.

Google’s own AI features have made this switch even more extreme, As they provide AI answers to people who were planning on using traditional search. The problem is that if things continue in this direction AI will eat itself. If news organizations and reporters can no longer earn a living by doing the hard work of researching an important story – that work just won’t be done – and chatbots will just have to resort to making up answers, relying on press releases or on propaganda – posted deliberately to mislead readers.

Content producers are obviously working to find a new structure where AI companies pay them for valuable information, but if that can’t, the open web may evolve into something very different.

So, let’s look at how AI is accelerating the collapse of the web’s information economy, what it means for the future of journalism and online content, and why the stakes are higher than most people realize. Before we dig into that let me tell you about this week’s video sponsor Tello mobile. If you're tired of overpaying for phone service, TM Mobile might be exactly what you need. With prices starting at just $5 a month, T lets you build your own phone plan, so you only pay for the minutes and data you actually use. Whether you're a student, a senior, or someone who's always on Wi-Fi, Tello gives you the flexibility to stay connected without breaking the bank. All plans come with unlimited texting, free mobile hotspot, and no contracts ever. And because Tello runs on a major nationwide network, you'll get great 5G, 4G, LTE coverage wherever you go. Need to switch things up? No problem. Tello’s month-to-month structure means that you can adjust your plan anytime, hassle-free. So, if you're looking for a smarter, more affordable way to stay connected, Tello Mobile offers flexible and affordable phone plans with prices up to $25 a month. Check out Tello using the link in my description.

The problem Google has with AI companies - is not the AI technology itself - but what that technology does to their existing business model. Google is not really in the business of giving you the best answer to your search questions; it's in the business of giving you a good enough answer so that you stick with google search, while encouraging you to click on an ad. If you ask what’s the best webcam, Google has to ensure that you will see ad placements paid by its customers at the top of the list. Google is now pitching that their new AI driven approach to search allows you to “let Google do the Googling for you.”

In theory, with the AI mode you can still see the links and use your own judgment, but not many users are clicking on the links anymore. And because their advertiser driven business model needs to continue working, the answers you get - might be colored by the need to sell certain products.

If you ask for health advice – as it appears many users are – will you be served information from reputable sources or from questionable supplement sellers who are paying for clicks?

I’m not sure how much you can trust this AI stuff…Recently I have been getting all these adverts on YouTube from a guy who says he is 68 – he’s also AI – and he has been doing Tai Chi and is in great shape. And I figure I’d like to be AI and in great shape too – but when I click on the link it tries to get me to do those exercises that you see old people doing at the park – and I don’t want to do that – as I worry that they’ll scramble my brain – like they did for this guy. I mean look at what he wrote on the white board… I started Tai Chi Chi fifty. Now I now I’m 68 Doctors ack ask me for advice.

They’re probably not ack asking him for advice on how to write… Anyhow…

Good journalism is expensive to produce. It requires reporters on the ground, editors with judgment, and teams of fact-checkers. Traditionally, this work was funded through either subscriptions or advertising. If users stop visiting websites, the incentive to produce original content disappears entirely. Why would anyone write a news report - just for it to be scraped by a dozen AI bots, scrambled up with other news and delivered to an audience – who have no interest in how it was created? The result would be fewer investigations, fewer foreign correspondents, fewer deep dives into complex issues. The web turns into a hall of mirrors—reflecting summaries of summaries, AI hallucinations and press releases, with no original source in sight.

This affects publishers, advertisers, and AI companies alike. AI tools depend on fresh, high-quality content so that they can function. They need journalists to gather facts, analysts to interpret data, and reviewers to test products. If those people can’t earn a living, the source material dries up. The bots will still generate answers—but those answers will be stitched together from outdated articles, press releases, and propaganda. AI ends up eating itself.

Publishers are trying to fight back—through licensing deals, lawsuits, and technical blocks—but the economics are brutal. If the ad money flows elsewhere, and the traffic never arrives, the open web may not survive in its current form. AI won’t just disrupt journalism. It will hollow out the information ecosystem it was trained on.

Journalism isn’t the only thing under threat. The entire ecosystem of online reviews—once a vital part of consumer decision-making—is breaking down too.

Before the internet, reviews were published in magazines. But those weren’t always trustworthy. Big advertisers could influence editorial decisions, and readers had limited ways to verify claims. The internet offered an alternative: independent reviewers who built trust through transparency and consistency. Tech reviewers, car reviewers, and niche experts earned loyal audiences by being honest—and monetized that trust through views and subscriptions.

Online reviews were one of the best things about the early internet. Platforms like Amazon offered a new kind of transparency—real feedback from real users. For shoppers, this was a breakthrough. You didn’t have to trust an advert or a sales pitch. You could read what other consumers thought.

That trust didn’t last. Sellers began gaming the system. Some offered free gifts or discounts in exchange for five-star reviews. Others paid bot farms to flood platforms with fake praise—or to sabotage competitors. Studies show that in categories like electronics and supplements, the majority of reviews may be fake.

Now AI tools are summarizing reviews—but they don’t distinguish between honest feedback and manipulated content. If trusted reviewers lose traffic because their work is scraped and repackaged without attribution or compensation, their incentive to produce high-quality, unbiased reviews disappears.

The review ecosystem isn’t just about choosing the right headphones or booking a good hotel. It’s part of how we evaluate truth. If AI tools can’t be trusted on this front—because their source material is compromised—they lose a key pillar of credibility.

As AI tools scrape and summarize content – often without permission or compensation, some creators are pushing back.

The musician Benn Jordan – who has one of the most interesting channels on YouTube - developed a tool called Poisonify to protect artists from unauthorized AI training. It’s designed to prevent AI companies from using music without the creator’s permission.

His video on this project – along with his most recent one on AI companies that are logging and sharing your vehicle's location without your consent are well worth watching. I’ll include links in the description.

Poisonify works by adding imperceptible “adversarial noise” to a music track—so the audio sounds unchanged to human listeners but sounds entirely different to AI scrapers. His software not only prevents AI’s from replicating an artist’s music without permission, but it may even punish them for taking the music without permission by corrupting their training data. If scraped, the poisoned tracks can possibly degrade an AI model’s performance, making it harder for AI companies to profit from stolen content.

Publishers are deploying countermeasures too. Cloudflare and other infrastructure providers now offer tools to block AI crawlers. Some media groups are negotiating licensing deals with AI firms and others are suing. The New York Times, for example, has filed lawsuits against OpenAI and Microsoft, arguing that their models were trained on copyrighted journalism without permission.

The most effective strategy may be branding. In a world where AI can mimic tone and summarize content, personality becomes a moat. Publishers are promoting individual voices—columnists, YouTubers, Substack writers—as a way to build loyalty and retain traffic. The Wall Street Journal recently advertised for a “Talent Coach” to help journalists build personal brands. This is based on the idea that readers follow people, rather than platforms.

This shift mirrors the rise of the creator economy, where independent journalists, analysts, and reviewers build direct relationships with their audiences—through newsletters, podcasts, and paid subscriptions. But even these models depend on visibility. If AI tools intercept the audience before they reach the creator, the economics still collapse.

The irony is that this strategy—building personality as a moat—only works if the personality is real. AI-generated influencers are already gaining traction. Synthetic voices, synthetic faces and synthetic opinions. If the last defensible asset is authenticity, then the next arms race is over what it means to be real.

Some startups are trying to build new models around AI access. Tollbit, which describes itself as a “paywall for bots,” allows content sites to charge AI crawlers variable rates—charging more for new stories than old ones. The firm argues that charging for access incentivizes uniqueness, unlike traditional search which rewards generic content.

Another approach comes from ProRata, which proposes that ad revenue from AI-generated answers should be redistributed to the sites whose content contributed to those answers. Its engine already shares revenue with over 500 partner publications.

A growing concern with users turning to AI chatbots for information is that the companies behind them have been aggressively overselling their capabilities. These systems are pitched as superintelligent beings—smarter than any expert, unbiased by design, and capable of answering anything. That framing gives users a false sense of confidence in the output. Elon Musk has claimed that his chatbot Grok is more intelligent than PhD holders in every discipline, that it can discover new physics, and outperform humans on Humanity’s Last Exam—a benchmark designed to test reasoning across 100 disciplines. Of course, he has also been claiming that his cars can drive themselves for nearly a decade – so we should probably take these claims with a pinch of salt.

Angela Collier made an excellent – and funny - video a few weeks ago called “Vibe Physics”, where she explains how tech entrepreneurs like the guys on the All In podcast mistake chatbot fluency for intelligence—especially when the bot starts talking about subjects outside their expertise, like her field: physics. They’re quick to spot errors when the chatbot discusses something they know, but once it shifts to unfamiliar territory, they’re suddenly amazed by its brilliance. The illusion of superintelligence kicks in precisely when the user lacks the tools to evaluate the answer.

Meta – which really ought to go back to calling itself Facebook - says that its AI-powered nerd glasses will give users a cognitive edge, and who knows – they might. Maybe trying to ignore the adverts being served to you all day long while you are trying to focus on more important things will boost your cognitive abilities.

When users stop reading news outlets—where they understand the editorial standards, reputations, and biases of both the journalist and the publisher—and start asking chatbots for answers assuming neutrality where bias may exist, they’re not just changing platforms. They’re changing who they trust to explain the world.

And that trust may be misplaced.

AI systems are not neutral. They reflect the data they’re trained on, the assumptions of their coders, and the incentives of the companies that deploy them. Bias can creep in through training data, algorithmic design, and moderation choices. Sometimes it’s accidental – but sometimes it’s deliberate.

Grok is a case study in what happens when those filters fail. After a July update designed to make it sound more “raw,” Grok began referring to itself as “MechaHitler.” Turkey banned the chatbot – not because of the Nazi stuff but because it also insulted President Erdogan.

Other bots have had breakdowns too - fabricating sources, hallucinating facts and misattributing quotes.

The myth of superintelligence encourages users to place their trust in machines that sound authoritative, even when they lack transparency or accountability. As chatbots become the default source of information, users may begin to rely on them more than qualified experts—especially in areas where expertise is hard-earned and context matters. This shift is happening at a time when public confidence in expert communities is already under pressure. According to Pew Research, trust in scientists has declined steadily over the last five years, with fewer than half of Americans now expressing strong confidence in scientific leaders. As AI systems take on the role of explainer-in-chief, the institutions that once anchored public understanding risk being sidelined. It is quite worrying that in that Economist chart I showed earlier that health information is most heavily impacted by the move to AI tools for information.

The open web was built on a simple exchange: publishers create content, users visit websites, and their attention is then monetized—funding the next round of reporting. AI tools are already breaking that loop.

If no one pays for the news, the news stops being reported. Investigations don’t happen. Deep dives into complex issues are replaced by AI summaries of press releases and hallucinations.

While some argue that citizen journalism can fill the gap, professional journalism can’t be replaced by individuals posting on social media. Eyewitness accounts and grassroots reporting can be powerful, but they don’t come with editors, legal teams, or the resources to spend months verifying leads and protecting sources. The BBC’s investigation into the Wagner Group in Libya, NBC’s exposé on forced adoptions in Christian boarding homes, and ProPublica’s reporting on U.S. Supreme Court ethics scandals all required institutional backing and sustained access to data and sources. These aren’t stories that could have been broken by someone with a smartphone and a Twitter account. Journalism isn’t just about being there—it’s about knowing what to do with what you find, and having the infrastructure to do it responsibly.

The disruption unfolding reaches beyond journalism and could reshape the entire information ecosystem. If the source material dries up due to lack of funding, the bots will still generate answers—but those answers will be stitched together from outdated articles, corporate PR, and propaganda. The web could begin to resemble a television camera pointed at a television—recycling recycled content.

I have talked in the past on this channel about the strategy of blitzscaling - where well-funded tech disruptors prioritize growth over profitability. The idea is that rapid expansion attracts high market valuations, allowing venture capitalists to flip unprofitable (or barely profitable) companies to the public before they ever find a sustainable source of profit. Blitzscaled firms often lose money deliberately, undercutting incumbents with the goal of eliminating competition and raising prices later. Uber didn’t need to be profitable to reshape urban transport; it just needed to be fast, cheap, and relentless.

AI is now following the same playbook. Vast sums are being spent to build tools that have yet to find a viable business model. But even without a path to profitability, they could still destroy the economic scaffolding that supports journalism, education, and public knowledge.

Technology eventually comes for everything. But when it comes for the institutions that help us understand the world, the stakes are higher than most people realize. [Jesse Welles Clip]

Business models evolve. When the internet first appeared, it looked like newspapers were doomed—but the industry adapted, reinventing itself for the digital age. Napster made it seem like music would be free forever, threatening that industry. Yet musicians found new ways to earn—through touring, streaming, and direct fan engagement.

The information economy is facing a similar moment, but having access to high-quality information is essential to the functioning of the modern world. And while today’s disruption feels existential, it’s more likely that journalism and knowledge institutions will adapt to the new environment than vanish from it. The business models may change. The platforms may shift. But the demand for truth, context, and accountability isn’t going away.

If you found this video interesting, you should watch this video on Americas mortgage divide next. Don’t forget to check out our sponsor Tello using the link in the description. Have a great day and talk to you in the next video. Bye!