Ah, the fantastic AI that will change the world for the better… but how many casualties will it make along the way? At this very moment, millions of AI chatbots extract answers from articles available on the internet, causing immediate reading of that summary and diverting visits from the site or blog (and journalist or author) that actually worked the article.
This new model of information consumption is already destroying traffic and revenue for independent blogs and content sites that depend on visits to survive. AI platforms don’t compensate creators for the information they use, creating an imbalance in the digital ecosystem that could have irreversible consequences for online information diversity.
The Silent Drama of Independent Creators
Imagine what it’s like to spend years building a blog, feeding it with original content and winning loyal readers, only to suddenly see visits plummet. In the US, a personal finance blogger reported a 42% drop in traffic in just six months after Google started summarizing her tips in an AI snippet: readers got direct answers without clicking on the site, and advertising revenue fell 51% in the period (yes, because in the US, advertising revenue “it’s a thing”).
This scenario isn’t isolated: news sites and blogs worldwide have witnessed drastic drops in traffic and revenue since the rise of generative AI tools like ChatGPT, Bing Chat, Gemini, and Google’s new search results. In many cases, these chatbots “devour” third-party content and deliver it to users without them visiting the original source, resulting in zero traffic or revenue (of any kind) for the author.
The impact goes beyond numbers: each lost click represents not only vanishing revenue but also the breakdown of a direct relationship between creator and audience. When a user gets information through AI, the brand context is lost, the opportunity to discover other related content is lost, and mainly, the creator’s possibility to build a community around their work. That famous and hard-won #reputation that still makes the Shah of 5 tick.
It’s a structural change that some already view as an existential threat to the “open web” based on free content funded by advertising. Even more worrying: this revolution is happening in silence, away from media spotlights, affecting thousands of small authors who simply see their audiences disappear without fully understanding why.
The AI That Reproduces Others’ Content and Dries Up Site Traffic
Generative AI tools learn from billions of words and images available online, much of it from blogs and content sites published for free. In practice, these models can mimic information from these sites and instantly respond to users, often with the same substance they would find in an article or post, just without any click to the original site.
For example: Google launched the AI Overview function in its search engine, which shows a summarized text block at the top of the results page. This AI-generated summary provides users with the desired answer immediately and pushes traditional “blue” links down, drastically reducing the likelihood of them being opened. The result? Sites that previously led certain searches report massive audience losses. An analysis revealed that a page occupying first place on Google can lose about 79% of traffic on that topic when it appears below an AI summary.
The mechanics of digital cannibalization work like this: AI extracts the informational essence from multiple sources, recomposes it into an apparently original answer, and delivers a complete solution to the user. The user, satisfied with the immediate response, doesn’t feel the need to “dig deeper” by visiting the original sources. It’s the triumph of convenience over curiosity, synthesis over exploration.
The same effect is happening outside Google. Dedicated chatbots like ChatGPT or Perplexity AI answer user questions by consulting the knowledge they’ve aggregated (largely taken from online content), but rarely direct users to the sources. I myself am an example of this, as I do many searches through these “new” means. A study indicated that AI chatbots send 96% less referral traffic to news sites than traditional Google searches. In other words, almost no one leaves the chatbot environment to click on an original article, a complete reversal of the “open web” model, where search engines functioned as bridges connecting users to creators’ pages.
Real Cases: When Numbers Tell Human Stories
Not by chance, reports of visit collapses on independent blogs and sites are multiplying. Owners of small niche sites (from cooking to DIY tutorials) have seen visitor numbers shrink abruptly in recent months. An illustrative case: a digital entrepreneur in the US watched her home projects site traffic drop 70% from one month to the next, leading to a 65% drop in advertising revenue, a loss of tens of thousands of dollars. “Google now scours my tutorials and displays the step-by-step directly in search, and readers no longer need to come to my blog,” lamented this creator.
Behind each percentage is a personal story: the mother who created a recipe blog to supplement the family budget and now sees the extra income disappear; the tech specialist who built an online reputation over years and finds himself replaced by generic answers; the freelance journalist who depended on organic traffic to sustain independent investigations.
The phenomenon repeats across numerous sectors: travel blogs (where detailed guides are replaced by automatic itineraries), technology (where in-depth analyses/reviews compete with instant summaries), cooking (where family recipes become algorithms), personal finance (where personalized advice is generalized), all see their audience diminish as AI offers quick, pre-chewed answers.
For larger sites, the trend is also worrying. Since Google incorporated AI responses into search (SGE), the proportion of searches that don’t generate clicks has skyrocketed. A Similarweb survey showed that in the news sector, the no-click search rate jumped from 56% to nearly 69% between May 2024 and May 2025, after the AI Overview launch. In the same period, organic traffic to news sites fell from 2.3 billion to 1.7 billion visits, a gigantic drop representing hundreds of millions in lost advertising revenue.
Some online newspapers report losing half their Google traffic: according to Similarweb, sites like HuffPost and Washington Post had 50% fewer organic visits after the introduction of these automated responses. Even giants like the New York Times felt the impact, with the percentage of their traffic from searches dropping from 44% to 36.5% over the past three years.
The irony of the times: although tools like ChatGPT are referencing more journalistic content – ChatGPT mentions of news sites grew 25 times from 2024 to 2025 – this doesn’t compensate for the drop in visits via Google. The same analysis indicates that these new AI-driven accesses represent only a fraction of what was lost in organic clicks. In other words, the math doesn’t add up for creators: more mentions, fewer visitors, revenue in free fall.
Without Creators, AI Also Runs Out of Content
This entire dynamic exposes a fundamental paradox that could determine the future of the internet: generative AI thrives by feeding on content produced by humans, but does so in a way that undermines the sustainability of those same producers. AI doesn’t compensate or credit the vast majority of sources it uses for training or immediate reference. Blogs, articles, and posts become free fuel for large models, while their authors lose the audience and compensation needed to continue creating.
In the short term, this is a unilateral advantage for Big Tech, which reaps the fruits of others’ content without paying. But long term? The disappearance of independent sites and traditional media would entail an “impoverishment of the internet and of AIs themselves.”
The math is simple and frightening: without new quality information produced by humans, AI models lose raw material for training. There’s already talk of the possibility of a “dead internet” dominated by automated content, in a vicious cycle where AI ends up regurgitating increasingly more material produced by itself. Researchers from Stanford and Rice universities warned in a recent study that repeatedly training AIs on data generated by AIs leads to progressive degradation of model quality and diversity, a phenomenon dubbed “model collapse.”
In short, without natural intelligence, there can be no artificial intelligence, or at least not with the richness and reliability we expect today. This paradox calls into question not only the financial viability of the “open web,” but the very sustainability of the digital knowledge ecosystem.
The model that prevailed over the past two decades, where millions of individuals and small businesses created content freely, funded by traffic and advertising, or voluntary subscriptions, is threatened. If the balance continues unbalanced, with AI extracting value without giving anything in return, many creators might close their doors. And if blogs, forums, specialized sites, and even media outlets shrink or disappear, the online knowledge ecosystem itself withers.
Ironically, AIs would be left without reliable sources to “drink” information from, harming the next generation of intelligent tools. We are, therefore, facing an unsustainable digital parasitism relationship: AI, as it stands, sucks nutrients from the “open web” but threatens to kill the host, and without a host, it also withers.
What Big Tech, Media, and Search Engines Are (Not) Doing
Faced with this emerging crisis, what has been the response from major tech companies and platform holders? So far, a worrying mix of denial, timid measures, and some legal disputes.
Google – whose dominant position in search makes it a central player – has downplayed the problem publicly. The company claims it continues to send “millions of clicks” to sites daily and hasn’t observed sharp traffic drops across the web as a whole. Google spokespeople have even classified studies indicating 70%–80% losses in organic clicks due to AI Overview as “flawed” and “unrepresentative.” In April 2025, confronted with reports from publishers devastated by traffic drops, a Google executive responded that it was “misleading to assign blame” specifically to AI, suggesting other factors might be at play.
This defensive stance ignores an evident reality: when a company with 92% market share in searches fundamentally changes how it presents results, the impact is systemic and immediate. Big Tech’s official position has been elegant denial: acknowledging that people are adopting AI search experiences, but not admitting direct responsibility for the publishers’ crisis.
Behind the scenes, however, Google knows the dilemma. CEO Sundar Pichai himself admitted that Mountain View ponders how the new AI search might “harm” traditional publishers. Still, the company moves forward, i.e., with Google’s business as usual, which now means integrating AI into everything, so as not to lose ground to competitors like Microsoft/OpenAI.
The bet is clear: keeping users in the Google ecosystem longer (reading generated responses instead of leaving for other sites) will be profitable through new ad formats integrated into AI responses. For Google, backing down doesn’t seem an option: in May 2025, it expanded SGE (Search Generative Experience) to all US users, despite protests from media associations that classified the appropriation of others’ content as equivalent to intellectual property theft.
Other tech players show no greater remorse. AI search startups and various applications that resort to generative models simply ignore the impact on creators, excited to offer instant convenience to users. Companies like OpenAI (ChatGPT creator) and Meta (which integrates conversational AIs into its platforms) have preferred to argue that AIs bring “new opportunities” for content discovery and enriching experiences, even though concrete data contradicts this technological optimism.
In August 2023, under pressure, OpenAI allowed sites to include a line in their robots.txt to prevent GPTBot from reading content, a minimal concession, since it places the burden on publishers and offers no financial compensation. It’s like offering a leaky umbrella during a great storm.
Media Reaction: Between Resistance and Forced Adaptation
On the traditional media side, a mixture of judicial resistance and desperate adaptation attempts is beginning to emerge. Several major communication groups have taken the legal route: in 2023, the New York Times filed a lawsuit against OpenAI, accusing it of using millions of copyrighted articles to train ChatGPT without permission. The case, still ongoing, could establish important precedents about copyright in the AI era (I didn’t find anything about this on Portuguese sites, if anyone knows, send it).
Other publishers, like the Associated Press, chose to negotiate licensing agreements: AP closed a deal to allow OpenAI to train models on its archives, possibly in exchange for some compensation. However, so far these are exceptions that confirm the rule: most creators don’t have negotiating power to extract similar deals.
Most news sites try to protect themselves by erecting paywalls – closed content requires login or subscription, which makes life difficult for AI scrapers, but also reduces visibility in search engines. It’s a defensive strategy that might work short-term but creates the risk of digital isolation: paid sites might survive financially but lose cultural relevance.
Some outlets, like Reuters, NY Post, or Business Insider, managed to gain referral traffic from ChatGPT (Reuters grew ~9%, for example), possibly for being sources that AI itself highlights prominently. But even that isn’t guaranteed or significant: the New York Times, despite being among the most referenced sites by ChatGPT, had a meager 3.1% gain in chatbot visits – while simultaneously seeing part of its organic traffic crumble. In other words, even when AI “helps,” it doesn’t even come close to paying the bills.
Regulation (…!)
And regulators and governments, are they paying attention? So far, signs of a systemic solution are tenuous and discouraging. In the European Union, some form of copyright content protection and transparency about training data is discussed under the AI Act, but nothing concrete regarding creator compensation or revenue-sharing mechanisms.
In the UK, media groups filed complaints with competition authorities, arguing that Google might be abusing its dominant position by introducing AI Overview and suffocating external traffic. The investigation is ongoing, but precedents aren’t encouraging: tech regulators have historically been slow, and Big Tech, quick to adapt or circumvent new rules.
In the US, conversely, there are political forces blocking any regulatory initiative: at the end of 2024, legislators included a clause in a fiscal law preventing states from creating new AI regulations for 10 years, reflecting lobbying that equates limiting AI to “giving China an advantage.” Big Tech, of course, supports this line. Figures like Nick Clegg (Meta executive) defend that publishers should be able to opt out of providing content for training, but warn that making this mandatory would “kill the AI industry” in countries that adopted it, an argument that reduces hope for voluntary compensation measures.
In short, regulation is behind and, in some cases, captured by pro-AI narratives that defend technological competitiveness over creative sustainability. There is, to date, no fair compensation model in force: no global fund to pay creators for content used, no tax imposed on AI companies, no clear revenue-sharing mechanism.
What we see are isolated and insufficient movements: companies like Reddit and Stack Exchange started charging for API access (to discourage mass data extraction), and artist and writer communities pressure for rights over data used in training. However, for millions of bloggers, independent journalists, and niche creators, the feeling is one of total helplessness. Many don’t have individual power to negotiate with giants, nor specific legal protection, and see the traffic and revenue “tap” closing before their eyes while awaiting regulation that may never come.
Surviving in the AI Era: Strategies for Content Creators
Faced with this scenario, independent creators and small online publications seek real alternatives to reinvent themselves. There’s no silver bullet, but various strategies are being tested – from radical business model changes to leveraging AIs themselves as unlikely allies. Here are some possible paths, each with its merits and limitations:
1. Focus on Paid Newsletters and Direct Subscriptions
One option is building a base of loyal readers willing to pay for content, instead of depending on anonymous traffic from search engines. Newsletter platforms like Substack, ConvertKit, or Revue (or even good old email marketing from the old guard) allow journalists, specialists, or bloggers to directly monetize their most dedicated audience.
The logic is simple: if Google won’t send readers, you need to go get them directly in their inboxes. Thus, even if organic visits drop, subscriber revenue keeps the business afloat. Several Portuguese and international creators already follow this model, offering exclusive content, in-depth analyses, or early access via newsletter for a small monthly fee.
It’s a return to an almost artisanal model of digital patronage, reminiscent of old newspaper subscriptions, but it can work for niches with highly engaged audiences. The challenge: it requires a mental shift both from the creator (who must think in premium value terms) and the audience (who must be willing to pay for what was previously free). And both, in the Portuguese case, are tremendously difficult.
2. Closed Communities and Strategic Paywalls
Along this line, some opt to create private communities – whether on Discord, Slack, exclusive forums, or closed social media groups – where members pay to access content and direct interaction with the creator. Content sites can implement paywalls (partial or total) to ensure that only those who contribute financially access certain articles.
While this might reduce initial reach, it at least guarantees that when content is consumed, there’s a financial return. Additionally, content behind login makes indexing by AI bots difficult, preserving some control over how information is used.
Member-only type models can also strengthen readers’ sense of community and belonging, something a generic AI doesn’t offer. Human interaction, debate, the possibility to influence editorial direction – these are intangible values that justify payment for many users.
3. Diversify Platforms: Video, Podcasts, and Social Media
“If you can’t beat them…” Many creators are migrating part of their content to formats less affected by textual AI. YouTube videos, podcasts, or even content on networks like TikTok and Instagram Stories represent territories where generative AI doesn’t yet completely dominate.
On one hand, these platforms have their own search algorithms; on the other, it’s harder for a textual chatbot to replace the experience of watching an original video, listening to a podcast conversation, or following a story’s evolution through sequential posts.
There are blogs that only worked with text that transformed into YouTube channels or added a weekly podcast, ways to reach audiences where they are and diversify revenue sources (like video advertising, sponsorships, merchandising, Super Chats, etc.). As cited in an analysis, creators who detected the Google drop now invest in YouTube tutorials and webinars, channels that have converted audiences better than old organic search. But even that starts to be a world full… of nothing.
The counterpoint: this strategy requires different skills (video editing, podcast techniques, understanding social media algorithms) and can dilute the creator’s identity across multiple platforms.
4. Embrace AI as a Tool, Not as an Enemy
Another counter-intuitive alternative is using AI in the creator’s favor. This can mean anything from employing writing assistants (like ChatGPT itself or the now-famous Claude) to gain productivity in article elaboration, to creating chatbots or interactive experiences based on the site’s own content. It requires spending dozens of hours learning, but nothing is done “without a degree,” as our dads used to tell us.
For example, a blog can train an AI model with its articles and offer readers a “virtual assistant” on the site. This way, users who like ChatGPT-style conversation at least do it within the creator’s page and not on Google. Some media explore AIs to personalize content recommendations to subscribers, or to generate utilitarian content (like summaries, puzzles, translations) that add value to the offering.
Evidently, there’s a delicate balance: embracing AI doesn’t mean surrendering the editorial line to it, but rather automating mechanical tasks and focusing human creativity where it makes a difference. It can also involve optimizing content for AIs themselves: the so-called SEO for AI or AEO (Answer Engine Optimization) to structure information so models correctly cite the brand or recommend the creator’s service.
It’s counter-intuitive, but some experts believe that “if you can’t prevent AI from using your content, at least teach it to point back to you.” The strategy involves creating content so valuable and well-structured that AI itself feels compelled to cite the source.
5. Ultra-Niche Specialization and Inimitable Expertise
An emerging trend is hyper-specialization in areas where AI still can’t compete effectively: unique personal experiences, analyses based on exclusive contacts, original research requiring physical access or complex human relationships. But Portugal has always been a ridiculously small market for this type of possibility. But anyway, let’s go with it:
The idea is simple: become so specific and authoritative in an area that AI, however sophisticated, can’t replicate the depth of knowledge. A wine blog that includes vineyard visits, producer interviews, and in-person tastings offers value no algorithm can match. Similarly, investigative journalism that depends on confidential sources, or technical analyses requiring hands-on practical experience.
6. The Hybrid Model: Free + Premium
Many creators are adopting a hybrid model: keeping part of the content free (for discovery and SEO), but reserving the best analyses, exclusive content, or early access for paid subscribers. It’s an attempt to have the best of both worlds: visibility through free content and sustainability through premium.
This strategy works especially well for creators who can establish clear value differentiation between what they offer for free and what they reserve for paying customers. Free content serves as “bait” to demonstrate quality and attract potential subscribers.
7. Reevaluate the Career… (oh, right)
Finally, some, in a sarcastic but not unrealistic tone, suggest that content producers consider other professional paths less threatened by automation. After all, in a world where even creative writing is imitable, maybe it’s worth more to be a maintenance technician, plumber, or mechanical turner, tangible jobs that an algorithm can’t execute remotely (at least for now).
“Jokes aside,” the frustration is real: many bloggers and independent writers feel the rug has been pulled from under them and seriously consider career changes. For new aspirants, the idea of starting a blog today, under current conditions, might sound so thankless that they simply give up before trying. It’s a loss of talent and creative diversity that impoverishes the internet and may be the true hidden cost of this technological revolution.
Is There Still a Future for Those Who Create Online Content?
The situation is horrifying for online content creators. Traffic in free fall, destroyed business models, tech giants innovating in apparently indifferent ways to the livelihood of millions of people who helped build the web as we know it. Is this the end of the golden age of blogs and the “open internet”? Is the future reserved only for AI-generated content, replicated ad infinitum, while human voices fall silent?
It’s important to note that internet history is made of disruptive cycles. “The death of blogs” was proclaimed before with the rise of social media, and many survived by finding new niches and formats. The arrival of generative AI is, undoubtedly, a major shake-up, however, not necessarily a total apocalypse.
Some types of content and human approaches will continue to have value that AI can’t easily replicate: investigative journalism, specialized opinion based on real experience, humor and genuine empathy, authentic personal experiences, creativity, analyses that connect apparently disconnected dots. These human qualities tend to stand out on a web saturated with generic automated material.
Readers and users might come to increasingly value authenticity when bombarded by AI texts. There are signs that more sophisticated audiences are already beginning to distinguish between human and artificial content, preferring the former for important issues or when seeking genuinely original perspectives.
Additionally, market and social pressures might force structural adjustments. If AI response quality begins to degrade due to lack of new quality data (remember the “dead internet” risk), AI companies will have economic incentive to invest in original content production and, who knows, directly funding certain creators or establishing fairer partnerships… let’s hope we’re still alive by then.
The European Union, with its history of tech regulation, might lead initiatives forcing Big Tech to share revenues with content creators… but that’s asking too much from those who have already shown they’re being fed by interests that aren’t European.
Ultimately, if users understand the value of supporting creators, whether by paying for content or preferring original sources instead of mere syntheses, this can profoundly influence future trends.
There are growing movements of “slow web” and “digital minimalism” that value quality over quantity, depth over convenience.
Future Scenarios: Three Possible Paths
Scenario 1: The Great Consolidation In this future, only the biggest players survive by negotiating licensing deals with AI companies. Result? Small creators disappear or are absorbed by platforms that offer them crumbs. The web becomes more homogeneous, controlled by few voices, but financially sustainable for those who achieve scale.
Scenario 2: The Micropayment Revolution Emerging technologies (blockchain, instant micropayment systems) allow users to automatically compensate creators for each AI query that uses their content. An ecosystem develops where AI pays small amounts whenever it cites a source, creating a sustainable value-sharing model.
Scenario 3: The Renaissance of Authenticity Artificial content saturation generates a contrary reaction. Audiences increasingly seek authentic human experiences, small but connected communities, content reflecting unique perspectives. Creators specialize in ultra-specific niches and prosper through direct connections with their audiences.
In Sum, Adaptation is Fundamental
The future of online creators is tremendously challenging, but not necessarily nonexistent. There will be natural consolidation: those who can adapt, innovate in format, and nurture loyal communities will have better chances of prospering in a world with omnipresent AI. Many will have to combine various strategies – a bit of paywall, a bit of SEO for AI, platform diversification, closed communities – to piece together their “digital livelihood.”
Others, unfortunately, won’t resist – just as many newspapers didn’t resist the digital transition of recent decades, or as many physical stores didn’t survive e-commerce. The “open web” as an ideal faces perhaps its greatest survival test, but this doesn’t necessarily mean its death – it might even mean its evolution into something different and, potentially, more valuable.
What was lost in quantity might be compensated in quality. If AI takes care of generic and utilitarian content, humanoids (!) can focus on what they do best: create emotional connections, offer unique perspectives, investigate what no one else investigates, entertain and inspire in ways no algorithm can replicate.
Collective Responsibility
AI companies must recognize they can’t prosper by indefinitely stealing from an ecosystem without giving anything back. Revenue-sharing mechanisms, fair licensing agreements, or direct investment in original content creation aren’t just ethical issues, they’re long-term economic necessities.
Governments need to update regulations for a reality where algorithms consume intellectual property on an industrial scale. This might include taxes on data usage, compensation funds for creators, or transparency rules about sources used by AIs.
Digital platforms (Google, social media, etc.) must balance innovation with ecosystem sustainability. Offering tools only to then undermine users’ business models is a self-destructive strategy.
Creators themselves must embrace change and experiment with new models, instead of just lamenting the past. This requires humility to learn new skills, courage to experiment with different formats, and resilience to face financially difficult transition periods.
And content consumers – all of us – must recognize that quality information has a cost. If we value diverse perspectives, independent journalism, and original creativity, we need to be willing to support these values – whether through subscriptions, donations, or simply consciously choosing to visit original sites instead of always settling for automatic summaries.
Concluding, If Possible: The “Open Web” Didn’t Die, It’s Evolving
One unquestionable certainty remains: we continue to need original, relevant, and genuinely human content. If generative AI is here to stay, and everything indicates it is, let it be as a complement and amplifier of human creativity, not as its substitute.
There’s enormous intangible value in community, in direct dialogue with an author, in knowing the story behind a text, in following someone’s thought evolution over time. These are values no artificial intelligence can match, however sophisticated. The question isn’t whether there will be a future for content creators, but what type of future that will be.
As long as there are those who seek these “authentic human connections” and those willing to create with passion, rigor, and originality, there will be a future for online content. At least that’s what we want to believe, as creatives. It might not be a future equal to the past, possibly with fewer sites, smaller ones sustained by hybrid models and more engaged audiences, but it will be a future since that will happen anyway in whatever manner or format.
The health of the “open web” interests even those who today apparently exploit it without consequences: without creators, there’s no knowledge diversity, and without diversity, there’s no true innovation. Even the most advanced AIs depend on human creative richness to continue evolving and surprising.
It’s up to everyone – industry, governments, creators, and public – to decide whether this future will be rich in diverse perspectives or if we’ll accept an internet of easy but soulless answers. The “open web” isn’t dying; it’s being tested. And like in all difficult tests, those who survive will emerge stronger, more focused, and, hopefully, more valuable.
After all, in an era of instant and artificial information, scarcity isn’t in information – it’s in wisdom, perspective, humanity. And that, at least for now, remains an exclusively our monopoly. Until it stops being…
Sources and References
The information and data cited in this article were obtained from recent analyses and expert testimonies, including Similarweb and Pew Research traffic studies on the impact of Google AI Overview on click decline, articles from The Guardian and eWeek detailing significant traffic drops on sites after AI response introduction, as well as reports from Exame and Fast Company Brasil about audience migration to chatbots and the consequent decrease in organic visits and advertising revenue.
Also referenced were specialist reflections on the paradox of dependence between AI and content creators and News/Media Alliance comments criticizing content appropriation by generative models.
Finally, market initiatives and reactions are mentioned, from lawsuits (like NY Times against OpenAI) to adaptation strategies by independent creators, as discussed in forums and articles aimed at the blogging and digital marketing community. All citations and references are linked to respective sources for consultation and verification.
Thank you for your patience if you made it this far.
