The Great Unlinking: How AI Overviews Defund the Journalism They Devour

On 4 February 2026, the Ahrefs content team published a single chart that should have been treated like a public health alert. It showed that when an AI Overview appears at the top of a Google search results page, the top-ranked organic link beneath it now receives 58 per cent fewer clicks than the same page would have received before AI Overviews existed. In April 2025, the same analysis had measured the decline at 34.5 per cent. In nine months, a feature initially described as an enhancement to search has roughly doubled the amount of traffic it diverts from the websites whose content it summarises. The 300,000 keywords Ryan Law and Xibeijia Guan analysed were not edge cases. They were the queries the open web has historically depended on for its survival.

That chart did not make front pages. AI Overviews did.

The numbers it represents arrive at a peculiar moment. In late January 2026, individual publishers logging into their AdSense dashboards discovered overnight earnings had collapsed by anything between fifty and ninety per cent. Daily revenue of five hundred dollars fell to thirty-five. Country-level coverage dropped to figures that read like a postscript: Germany down sixty-four per cent, France sixty-three, Italy seventy-six, Spain ninety. Google later attributed the failure to third-party tag recognition issues in Ad Manager that cascaded through Ad Exchange. Whatever its proximate cause, the incident gave thousands of small publishers a glimpse of what life on the other side of the search economy now looks like. The trapdoor had not opened. It had merely groaned.

Press Gazette's May 2026 audit of the fifty most popular US news sites confirmed the broader pattern. Almost half had seen their year-on-year traffic in April fall by 20 per cent or more, despite a steady stream of breaking news. Forbes recorded the worst decline among the top fifty, down by nearly half. AP News had shed 46 per cent of its visits in a year. Athlon Sports lost 48 per cent. Some publishers grew, most notably Substack's newsletter network and Men's Journal, the latter quadrupling its traffic, but the direction of travel for the legacy news web was unmistakable: down, and accelerating.

This article is about that descent. It is also about who, in the end, gets to decide whether it continues.

The shape of the disappearance

Begin with the mechanics, because the mechanics matter. AI Overviews are the AI-generated summaries that now appear above traditional search results on Google for a growing share of queries. They are produced by Gemini, Google's family of large language models, drawing on a vast index of web content that includes, of course, the journalism that publishers have spent decades producing. The Overview answers the user's question directly. The links to the underlying sources remain, but they sit underneath an answer that has, by design, made clicking them unnecessary.

For two decades, this would have been considered a hostile redesign of the search results page. Search engines extracted value from the open web by indexing it, but they returned value through referrals: a query went in, a link came out, the publisher received a visitor who could be shown advertisements, persuaded to subscribe, or counted in the metrics that justified the salary of the reporter who wrote the piece. It was an implicit bargain rather than a contract, but it was durable. The bargain is now visibly breaking.

Cloudflare's data on the asymmetry between what AI companies take and what they give back is unsparing. In June 2025, Google's crawler scraped websites fourteen times for every referral it sent. OpenAI's crawler scraped 1,700 times for every referral. Anthropic's, the most extractive of the major systems, scraped 73,000 times for every visitor it sent on. By July 2025, Anthropic's ratio had risen to 38,000 pages crawled for each referred page visit, an imbalance that, as Cloudflare's Matthew Prince has argued repeatedly, is incompatible with the survival of the businesses that supply the underlying content. The Security Boulevard analysis published in April 2026 framed it crisply: large language models scrape publisher content thousands of times for every single referral they send back, destroying the advertising and subscription revenue that pays for the reporting being consumed.

There is something almost geological about the slowness with which this realisation has settled. Search engines have always taken more than they gave back in any narrow accounting; the value they created accrued elsewhere, in the form of an open web worth searching. AI search dispenses with that justification. The user gets the answer. The model gets the training data. The platform gets the advertising slot at the top of the page. The publisher gets a citation, sometimes, in a panel that the user is empirically unlikely to expand.

The case studies are not anecdotes

The temptation, when writing about systemic decline, is to reach for individual stories that humanise the abstraction. The trouble with the AI Overviews story is that the abstraction has already eaten the stories whole, and the bodies are not metaphorical.

The Planet D, a travel blog founded in 2008 by Dave Bouskill and Debra Corbeil, lost half its traffic in the months after Google launched AI Overviews in May 2024. Staff were laid off. Traffic then fell another 90 per cent. The site stopped publishing earlier in 2025. Charleston Crafted, a home improvement blog, lost 70 per cent of its traffic between March and May 2024 and saw a 65 per cent decrease in advertising revenue. Stereogum, one of the longest-running independent music publications on the open web, reported a 70 per cent collapse in ad revenue in 2025. Its founder, Scott Lapatine, attributed most of the damage to AI Overviews, though Facebook's and X's deprioritisation of links played a supporting role, and announced a transition to paid subscriptions in the hope of replacing what the platform had taken.

These are not boutique websites that misjudged the market. They are precisely the kind of mid-sized specialist publishers the early web was supposed to make possible: small enough to be intimate, large enough to be professional, dependent on advertising revenue that scaled with audience attention. AdExchanger, in its January 2026 reckoning, documented that publishers across the spectrum had lost between 20 and 90 per cent of their traffic and revenue as AI Overviews became the default mode of search. Business Insider's organic search traffic fell by 55 per cent between April 2022 and April 2025. HuffPost's desktop and mobile sites lost half their search referrals over the same period. The Reuters Institute's Digital News Project, in its 2026 predictions report led by Nic Newman, found global Google search traffic to news publishers had fallen by 33 per cent in 2025, with Google Discover down 21 per cent. The newsrooms surveyed expected a further drop of 43 per cent over the next three years. Only 38 per cent of news executives reported feeling confident about the year ahead, down from 60 per cent four years earlier.

The Reuters Institute's framing is worth quoting in spirit if not in length: 2026 is not the year that AI is coming for journalism. It is the year journalism's existing distribution layer has begun to dismantle itself in real time.

What is being defunded, exactly

Now consider what that distribution layer used to fund. The advertising and subscription revenue that has flowed through publisher websites paid for many things. It paid for celebrity gossip, listicles, sponsored content, and considerable quantities of search-optimised filler the open web will not, in itself, miss. But the same revenue stream also paid for the journalism no one else funds.

In the United States, the Medill State of Local News Report for 2025, led by Tim Franklin and informed by the foundational research of Penny Abernathy, found the number of news desert counties, those with no local news organisation at all, had risen to 213. Another 1,524 counties had a single remaining news source. Roughly 50 million Americans now live with limited or no access to local news. Newspaper closures continued at more than two a week, with the steepest losses concentrated in small, independently owned publications. Over two decades, the United States has lost nearly 3,500 newspapers and more than 270,000 newspaper jobs.

The numbers can be read in two ways. One is to note the local news crisis predates AI Overviews by a decade; print advertising's collapse and the dominance of social media did most of the damage first. The other reading, which is closer to the truth, is that the digital advertising economy that succeeded print was the lifeline allowing surviving local outlets and digital startups to make a partial recovery. Two-thirds of the more than 300 local news startups launched over the past five years are digital-only, and most depend on a combination of organic search traffic, advertising, and newsletter subscriptions. The decline in search referrals the Reuters Institute is tracking is not abstract for those outlets. It is the difference between an additional reporter and a wound-down operation.

Court reporting is a particularly clean example because the structure of the work makes the dependency visible. Covering a magistrates' court or a county court is labour-intensive, often unglamorous, and largely unprofitable except as part of a larger publishing operation whose other pages subsidise the public-interest reporting. When the operation's economic base erodes, court reporting is among the first beats to be cut, because no commercial entity is willing to pay directly for it. The Arizona Supreme Court's recent introduction of AI-generated summaries of rulings is a striking symbolic moment: a court system has begun automating the explanation of its own decisions because the human stenographers and beat reporters who once did the work are no longer reliably present in the room. The summaries will draw, inevitably, on the journalism that used to be written by those reporters, until that journalism, too, becomes scarce enough that the summaries begin to fail.

Health and science coverage is similarly load-bearing. During the pandemic, the role of science reporters in translating epidemiological evidence into public understanding was visible to anyone watching. Investigative reporting is even more concentrated. ProPublica, the Bureau of Investigative Journalism, regional non-profits and a handful of legacy newsrooms produce the bulk of accountability work in the English-speaking world. The economics of investigation are brutal: a single piece can take months and produce no traffic until it does. The cross-subsidy from high-volume, lower-effort content that finances the slow work is precisely what AI Overviews are dismantling. When the page about who the richest person in the world is no longer drives traffic to Forbes, the part of Forbes that does actual reporting becomes that much harder to sustain.

This is the load-bearing element the regulatory debate keeps gesturing at without quite saying. The damage from AI Overviews is not evenly distributed across content types. It is concentrated, by the logic of summarisation, on the pieces that can be summarised: definitional content, explainer journalism, listicles, evergreen reference material. The investigative scoop, the eyewitness reportage, the court transcript, the science explainer that took three weeks to get right: those are harder to extract, but they sit in publishing operations whose business model depends on the extractable pieces continuing to earn. The summary eats the canapés. The kitchen closes anyway.

The platforms' defence

Google's case for AI Overviews has been made most consistently by Sundar Pichai, who has argued in several settings that AI Overviews send users to a wider variety of websites than traditional search, and that publishers are misreading early data. At Google Cloud Next 2026 he sketched a future in which search becomes an agent management layer, with AI models interpreting queries, synthesising answers, and executing tasks across services. Asked about journalism in a June 2025 podcast with Lex Fridman, he said news and journalism would play an important role in the future, and that Google was committed to it.

The trouble with this defence is that it requires accepting the platform's metrics about its own behaviour. The Ahrefs methodology was deliberately constructed to control for the kind of measurement noise Google has previously invoked to explain away earlier declines. It compared 150,000 keywords that triggered AI Overviews against 150,000 informational-intent keywords that did not, using aggregated Google Search Console data covering the period before and after AI Overviews' widespread rollout. The 58 per cent decline is not a vibe. It is the result of one of the better-instrumented experiments the open web is capable of running on itself. And in February 2026, Penske Media Corporation, the publisher of Rolling Stone, Variety, Deadline, and The Hollywood Reporter, submitted that same Ahrefs analysis as part of its federal court memorandum opposing Google's motion to dismiss its antitrust lawsuit. The lawsuit, filed in September 2025, alleges Google has abused its search monopoly to compel publishers to accept AI summarisation of their content as the price of continued visibility in search. Penske's central argument is that the historic bargain, content for traffic, has been unilaterally rewritten and that publishers were given a choice that is no choice: leave Google search altogether, or accept the cannibalisation.

Google has moved to dismiss. Its position is that AI Overviews are summaries of information responsive to a user's query, not a separate product, and that displaying an Overview does not deprive users of alternatives. The same argument, more or less, is being made in Europe, where the European Publishers Council filed a formal antitrust complaint with the European Commission on 10 February 2026. The complaint, brought under Article 102 of the Treaty on the Functioning of the European Union, alleges Google's AI Overviews and AI Mode constitute an abuse of dominance: the dominant gatekeeper, in EPC chairman Christian Van Thillo's framing, is using its market power to take publishers' content without consent, without fair compensation, and without giving publishers a realistic way to protect their journalism. The EPC's membership reads like a roster of the European newsroom: DMG Media, The Guardian, News UK, Le Monde, El País, The New York Times. The European Commission had already announced an antitrust investigation into Google's use of publisher content for AI training in December 2025.

The platform's most consistent rhetorical move in response has been to insist the alternative to AI Overviews would be worse: a search experience that fails to keep pace with user expectations set by ChatGPT, Claude, Perplexity and other answer engines, all of which are themselves drawing on the same publisher content with even more extreme crawl-to-referral ratios. There is a real argument here, but it is also self-serving. The choice between AI Overviews and a competitor's worse extraction is a choice the platforms have set up for themselves. The choice the publishers are asking to make, which is whether their content should be used in AI summarisation at all without consent or remuneration, is the one the platforms have so far refused to offer in any meaningful form.

What the regulators have proposed

In January 2026 the UK Competition and Markets Authority, having already designated Google as having strategic market status in general search and search advertising in October 2025, proposed a set of conduct requirements that would force the platform to offer publishers a genuine opt-out from AI Overviews. The proposal is unusual in its directness: publishers would be able to withhold their content from AI Overviews and from the training of Google's broader generative AI services, including Gemini and Vertex, without losing visibility in traditional organic search. Google would also be required to ensure publisher content is properly attributed in AI results. The consultation closed on 25 February 2026. As of mid-May 2026, the CMA is reviewing responses and is expected to issue final conduct requirements in the coming months.

The opt-out, if implemented, would be the first time a major regulator has unbundled the historic implicit bargain at the level of explicit policy. Until now, the choice for publishers has been all-or-nothing: be in Google, accept whatever Google does with your content; or leave Google, lose most of your audience. The CMA's proposal would create a third option: stay in Google's index, but refuse the AI summarisation. The PPA, representing UK consumer magazine and B2B publishers, responded with cautious support. The News Media Alliance in the United States, led by Danielle Coffey, has called for similar interventions and described Google's late-January 2026 opt-out announcement as a welcome sign the company is starting to listen to publishers, while noting the gesture came only in response to sustained regulatory pressure.

There are good reasons to be cautious about what an opt-out actually achieves. A publisher that withdraws from AI Overviews and AI Mode loses presence in the surface Google is increasingly making the default. The competing AI search products, from OpenAI's SearchGPT to Perplexity to Anthropic's web-aware models, would not necessarily be covered by a Google-specific remedy. And the negotiating asymmetry between an individual publisher and a multi-billion-dollar platform remains stark, even with a regulator's hand on the scale. The European Publishers Council's complaint anticipates this and asks the Commission to go further: not just opt-outs but compulsory licensing, statutory remuneration, and structural separation of AI summarisation from the search interface.

The most interesting technical proposal has come, unexpectedly, from infrastructure. Cloudflare's Matthew Prince launched pay-per-crawl in private beta in July 2025, allowing website owners to charge AI crawlers a micropayment for each scrape. The platform sits at a useful chokepoint: roughly a fifth of the open web routes through Cloudflare in some form, which means a meaningful share of crawlers can be metered or blocked at the network layer rather than at the level of individual publisher policy. Pay-per-crawl assumes what regulators have been slow to acknowledge: the consent regime for AI training and AI summarisation is not, in any meaningful sense, opt-out. It is opt-in by silence, enforced by the absence of an enforcement mechanism.

A scenario worth taking seriously

Imagine, for a moment, the trajectory of a single mid-sized regional title under current conditions. The paper, a hypothetical composite of the kind described in the Medill report, employs twenty-two journalists across news, courts, council coverage, sport, and a small lifestyle desk. Its digital advertising revenue, the bulk of its income since print declined a decade ago, is roughly evenly split between Google AdSense and a direct-sold programmatic stack. Half of its traffic comes from Google search. By the end of 2024, AI Overviews had begun appearing on the kinds of queries that drove most of its evergreen traffic: how to register to vote, what time the local library opens, when the new school term starts, the names of councillors. By April 2025, the Ahrefs measurement at 34.5 per cent decline already meant a perceptible drop. By the time the February 2026 update lands and the figure climbs to 58 per cent, the paper has lost roughly a third of its overall digital traffic and close to forty per cent of its programmatic ad inventory.

Then, on 14 January 2026, AdSense earnings collapse for forty-eight hours. The technical fault is rectified, but the publisher's senior leadership, looking at the chart, understands they have just glimpsed the underlying volatility of their revenue base. The board commissions a review. By April 2026, when Press Gazette publishes its audit of the top 50 US sites, the paper has cut six positions: two court reporters, the local government beat reporter, a science writer who had been part-funded by a foundation grant, and two subeditors. Coverage of the magistrates' court reverts to police press releases. The council's licensing committee, previously covered by a reporter who knew the regulars, is now reported on, when at all, from agenda papers downloaded the morning after meetings.

This is not a thought experiment offered as melodrama. It is the rough operational shape of the choices being made, right now, in dozens of newsrooms. The Medill report's underlying finding, that closures and contractions are accelerating among small and mid-sized publishers, is not a function of AI Overviews alone. It is the product of compounded pressures: declining print circulation, social media de-prioritisation of links, programmatic advertising's collapsing yields, and now the redirection of search traffic to summarisation. The question is not whether journalism would have struggled without AI Overviews. It is whether AI Overviews are the policy choice that turns a difficult adjustment into an irreversible one.

Who decides

The question of who should decide how the value is distributed is the hardest one, and the one most likely to be answered by default rather than by design. Several candidates present themselves.

The first is the platforms. Google's stated position is that the existing bargain remains intact, that traffic patterns are simply shifting, and the company is adjusting its product to keep publishers visible. The series of updates announced in early 2026, including Further Exploration links and subscription labels, are real, but they are platform-administered concessions. Their existence depends on the platform's continued belief that they are necessary. As soon as the regulatory pressure abates, the architecture of the search results page is once again at the platform's discretion. The implicit governance is that whoever owns the surface decides the terms.

The second is governments and regulators. The UK CMA's strategic market status designation and proposed conduct requirements represent the most ambitious attempt yet to translate the implicit bargain into explicit policy. The European Commission, with the EPC's complaint now in its tray and a December 2025 investigation already running, has both the legal tools, in the form of the Digital Markets Act, and the political will. The US position is more fragmented: the Department of Justice has Google in court on separate antitrust grounds, and Penske's lawsuit is making its way through the federal courts, but congressional action on AI-specific competition policy remains largely aspirational. Regulators have the legitimacy to draw the line. The question is whether they can move quickly enough to matter, and whether opt-outs are a sufficient remedy or merely a way of formalising the existing power asymmetry.

The third is publishers themselves, acting collectively. The history here is not encouraging. Publishers have repeatedly failed to coordinate effectively against platform pricing power, partly because they compete with one another and partly because individual deals, of the kind Google and OpenAI have signed with selected outlets, fragment the bargaining unit. The European Publishers Council's complaint is a notable exception: a coalition action that names a structural problem rather than negotiating individual remunerations. The challenge is whether collective action can be organised at a global scale, given that the platforms operate globally and the publishers are dispersed across legal jurisdictions with different competition regimes.

The fourth is citizens. This is the candidate the policy debate has, so far, almost entirely avoided. The decision to redirect the economic value of journalism from the institutions that produce it to the platforms that summarise it has not been put to anyone. There has been no white paper, no green paper, no parliamentary debate framed around the question of what local accountability journalism is worth to a democracy and how its provision should be secured. The CMA's consultation is the closest thing to a public process and its remit is properly narrow, scoped to competition law rather than to the wider question of whether the architecture of information distribution should be a matter of private commercial discretion at all. The asymmetry between the scale of the decision and the smallness of the public forum in which it is being taken is, on any measure, striking.

A line that should be drawn

The position this article takes, after working through the data, is that the redirection of journalism's revenue base to AI summarisation is happening too quickly, on too large a scale, and with too little public deliberation for any reasonable observer to treat it as a market adjustment. It is a transfer of value. It is being effected by parties that did not produce the underlying content. The mechanisms by which it is occurring are, if not formally illegal, then certainly inconsistent with the bargains under which the content was produced. The regulatory and legal responses, in the UK, the EU, and through the Penske litigation in the United States, are appropriate and overdue. They should be supported, sharpened, and extended.

But the deeper point is that the question is not, ultimately, a competition law question alone. It is a democratic infrastructure question. The journalism being defunded is the journalism that makes local government legible, that holds courts and police accountable, that translates scientific findings into civic understanding, and that surfaces wrongdoing in time for it to be addressed. None of that is produced by the AI systems now distributing it. Some of it, the explainers and definitional content, can be reproduced after the fact by models trained on what previous journalism produced. The investigations, the eyewitness reportage, the long-cultivated source relationships, the appearance in court each week to take a note: those cannot be summarised because they have to be done first.

The communities that depend on that reporting, which is to say all communities, do not currently have a meaningful seat at the table where the value transfer is being decided. The first task of any serious policy response is to give them one. That means treating publisher opt-outs as a floor, not a ceiling; mandating compensation regimes for content used in AI summarisation; investing in public-interest journalism funds drawn from a levy on platform revenues, on the model some European jurisdictions have begun to consider; and, perhaps most importantly, naming the situation honestly. The web has not broken. It has been broken open, and someone with a basket is collecting what falls out.

What the journalists who produced this material would have wanted, had they been asked, is not the right framing, because most of them were not asked. What the readers who valued the reporting would have chosen is not the framing either, because they were not consulted. The decision is being made by the platforms that own the surfaces, the publishers who lack the leverage to refuse, and the regulators who are catching up. The 58 per cent figure Ahrefs published in February is a measurement of how much of the old settlement has already been cleared away. The questions that remain, about who gets to build what replaces it, and on whose terms, are still, just barely, open.

If they are to be answered in a way that preserves anything of the journalism the open web sustained, the conversation needs to happen now, in public, with the people who depend on the reporting in the room. The alternative, which is the trajectory already underway, is that the answer will be supplied by default, by the entity with the surface and the model and the advertising slot at the top of the page. That entity has already made its preferences clear. It will summarise what is left.

References

  1. Law, R. and Guan, X., “Update: AI Overviews Reduce Clicks by 58%“, Ahrefs Blog, 4 February 2026. https://ahrefs.com/blog/ai-overviews-reduce-clicks-update/
  2. “AI Overviews Reduce Clicks by 34.5%”, Ahrefs Blog, April 2025. https://ahrefs.com/blog/ai-overviews-reduce-clicks/
  3. Press Gazette, “Top 50 US news websites: Half of sites see traffic fall 20% or more in a year”, May 2026. https://pressgazette.co.uk/media-audience-and-business-data/media_metrics/most-popular-websites-news-us-monthly-3/
  4. AdExchanger, “The AI Search Reckoning Is Dismantling Open Web Traffic, And Publishers May Never Recover”, January 2026. https://www.adexchanger.com/publishers/the-ai-search-reckoning-is-dismantling-open-web-traffic-and-publishers-may-never-recover/
  5. Security Boulevard, “The AI Content Crisis: How LLMs Are Draining Media Revenue, and the Technologies Fighting Back”, April 2026. https://securityboulevard.com/2026/04/the-ai-content-crisis-how-llms-are-draining-media-revenue-and-the-technologies-fighting-back/
  6. ALM Corp, “AdSense Revenue Plunge January 2026: 90% Earnings Drop Hits Publishers Globally”, January 2026. https://almcorp.com/blog/adsense-revenue-plunge-january-2026-causes-solutions-recovery/
  7. PPC Land, “Google ad platform failures slash publisher revenue up to 90% overnight”, January 2026. https://ppc.land/google-ad-platform-failures-slash-publisher-revenue-up-to-90-overnight/
  8. European Publishers Council, “European Publishers Council files formal antitrust complaint against Google over AI Overviews and AI Mode”, 10 February 2026. https://www.epceurope.eu/post/european-publishers-council-files-formal-antitrust-complaint-against-google-over-ai-overviews-and-ai
  9. UK Competition and Markets Authority, “CMA proposes package of measures to improve Google search services in UK”, 28 January 2026. https://www.gov.uk/government/news/cma-proposes-package-of-measures-to-improve-google-search-services-in-uk
  10. UK Competition and Markets Authority, “CMA confirms Google has strategic market status in search services”, 10 October 2025. https://www.gov.uk/government/news/cma-confirms-google-has-strategic-market-status-in-search-services
  11. Press Gazette, “Google exploring updates to let publishers opt out of AI Overviews”, January 2026. https://pressgazette.co.uk/news/google-ai-overviews-search-cma-proposals/
  12. Newman, N., “Journalism, Media, and Technology Trends and Predictions 2026”, Reuters Institute for the Study of Journalism, University of Oxford, January 2026. https://reutersinstitute.politics.ox.ac.uk/journalism-media-and-technology-trends-and-predictions-2026
  13. Cloudflare, “The crawl before the fall, of referrals: understanding AI's impact on content providers”, Cloudflare Blog, 2025. https://blog.cloudflare.com/ai-search-crawl-refer-ratio-on-radar/
  14. Axios, “Penske Media sues Google over AI search overviews”, 14 September 2025. https://www.axios.com/2025/09/14/penske-media-sues-google-ai
  15. Columbia Journalism Review, “The Creative Approach Behind Penske's AI Lawsuit”, 2026. https://www.cjr.org/analysis/penske-ai-lawsuit-google-tactics.php
  16. Lapatine, S., “Getting Killed By AI”, Stereogum, November 2025. https://stereogum.com/2478838/stereogum-relaunch/tcb
  17. Medill School of Journalism, Northwestern University, “News deserts hit new high and 50 million have limited access to local news, study finds”, State of Local News Report 2025, October 2025. https://www.medill.northwestern.edu/news/2025/news-deserts-hit-new-high-and-50-million-have-limited-access-to-local-news-study-finds.html
  18. News/Media Alliance, “Statement on Google AI Mode Opt-Out Announcement: Cautiously Optimistic”, January 2026. https://www.newsmediaalliance.org/nma-statement-on-google-opt-out-announcement/
  19. TechCrunch, “Cloudflare launches a marketplace that lets websites charge AI bots for scraping”, 1 July 2025. https://techcrunch.com/2025/07/01/cloudflare-launches-a-marketplace-that-lets-websites-charge-ai-bots-for-scraping/
  20. Press Gazette, “Global publisher Google traffic dropped by a third in 2025”, 2026. https://pressgazette.co.uk/media-audience-and-business-data/google-traffic-down-2025-trends-report-2026/

Tim Green

Tim Green UK-based Systems Theorist & Independent Technology Writer

Tim explores the intersections of artificial intelligence, decentralised cognition, and posthuman ethics. His work, published at smarterarticles.co.uk, challenges dominant narratives of technological progress while proposing interdisciplinary frameworks for collective intelligence and digital stewardship.

His writing has been featured on Ground News and shared by independent researchers across both academic and technological communities.

ORCID: 0009-0002-0156-9795 Email: tim@smarterarticles.co.uk

Listen to the free weekly SmarterArticles Podcast

Discuss...