Is AI Slop Killing the Internet?

Key Concepts

AI Chatbots vs. Search Engines: The shift from traditional search engines to AI chatbots for information retrieval.
Information Economy Disruption: How AI tools are altering the economic models of content creators, particularly news providers.
Monetization Erosion: The decline in revenue for publishers due to AI tools providing direct answers instead of links.
Loss of Visibility: The significant drop in search result visibility for news organizations.
Structural Shift: The fundamental change in how users access information, bypassing original sources.
Accountability Gap: AI systems lack the accountability of human-created content.
"AI Eating Itself": The potential for AI to degrade the quality of information if original content creation is no longer sustainable.
Creator Economy: The rise of independent creators and their reliance on audience engagement.
Branding and Personality as Moats: Using individual voices and brands to build loyalty in the face of AI mimicry.
Adversarial Noise: A technique to protect digital content from unauthorized AI training.
Blitzscaling: A business strategy prioritizing rapid growth over profitability, often at the expense of incumbents.
Trust and Bias in AI: The challenges of AI neutrality, potential for bias, and the myth of superintelligence.
Ecosystem Collapse: The potential for AI to hollow out the entire information ecosystem.

AI's Transformation of Information Access and its Economic Impact

Artificial intelligence (AI) tools are rapidly changing how people find information online, a transformation occurring faster than publishers can adapt. Users are increasingly turning to AI chatbots like ChatGPT, Claude, and Perplexity for answers, bypassing traditional search engines. This shift fundamentally alters the internet's economic landscape, particularly for news providers who invest in gathering information that they can no longer monetize effectively.

The Shift to AI Chatbots

Over the past few years, millions of users have migrated from search engines to AI chat tools for research, recommendations, and immediate answers. These tools directly provide answers, rather than links to primary sources that users previously trusted. This trend leads users to place increasing faith in AI systems, which are trained to mimic authority by scraping the web for information but lack accountability.

Google's AI Integration and Publisher Impact

Google's introduction of "AI Overviews" and "AI Mode" directly answers user questions on search result pages, often without crediting original sources. While this offers a seamless experience for users, eliminating the need to refine search terms or consult multiple articles, it has been "disastrous" for publishers.

Data and Statistics:

A report by Enders Analysis, using Sistrix data, indicates a collapse in news visibility in search results.
The Mirror's presence on Google has decreased by 80% since 2019.
The Mail has lost over half of its visibility.
The Financial Times experienced a 21% drop in traffic in the spring.
This decline is attributed primarily to Google's AI features, not publisher strategies.

Structural Economic Erosion

This shift is described as "structural" because AI tools intercept audiences before they reach the original information sources. Google's AI Overviews affect even users not actively seeking AI answers. Publishers cannot opt out without disappearing from search entirely, leading to the rapid erosion of the economic model that sustained journalism—clicks, subscriptions, and advertising—as AI systems extract value without returning any.

Broader Impact:

The Economist chart indicates that news and media are not the only affected sectors. Science, education, and reference sites (like Wikipedia) are also experiencing significant declines.
Health information is identified as the most impacted area.

Measurable User Behavior Changes

Search engines, once the primary gateway to the internet, are seeing their influence wane.

In April, Apple reported the first-ever decline in Safari search volume, attributed to users shifting to AI chat tools.
A December TechRadar survey found that 27% of U.S. users and 13% of UK users now use AI tools instead of search engines for information gathering, citing speed, specificity, and ease of use.
Google referrals to news sites have dropped from approximately 65% in 2019 to 30% currently, though this decline predates AI Overviews.

The Fading Discoverability of Information

The notion that "nothing ever disappears from the internet" is being challenged. Information can become effectively lost if it becomes impossible to find, even if never erased. Google's declining efficiency in finding information is attributed to web saturation with SEO-optimized content and an increase in sponsored search results, forcing users to sift through clutter. AI tools emerged as internet search was faltering, with The Atlantic reporting that AI chatbots have already replaced search for over 25% of Americans.

The "AI Eating Itself" Scenario

Google's own AI features exacerbate this shift by providing AI answers to users intending to use traditional search. The concern is that if this trend continues, AI will "eat itself." If news organizations and reporters can no longer earn a living through investigative work, that work may cease to be done. Chatbots would then be forced to fabricate answers, rely on press releases, or disseminate propaganda, leading to a decline in the quality and integrity of information.

Publisher Responses and the Future of the Open Web

Content producers are attempting to establish new models where AI companies compensate them for valuable information. However, if this fails, the open web may evolve significantly. The current situation accelerates the collapse of the web's information economy, posing a threat to journalism and online content.

Google's Business Model and AI's Challenge

Google's core business is not solely about providing the best search answers but about offering "good enough" answers to retain users and encourage ad clicks. Their new AI-driven approach, "let Google do the Googling for you," aims to streamline this. While AI Mode theoretically allows users to still see links, click-through rates have diminished. This advertiser-driven model raises concerns that AI-generated answers might be influenced by the need to promote specific products, particularly in sensitive areas like health advice.

The Cost of Quality Journalism and the Threat to Original Content

Producing high-quality journalism is expensive, requiring reporters, editors, and fact-checkers. Traditionally funded by subscriptions or advertising, this work loses its incentive if users stop visiting websites. The prospect of news reports being scraped, aggregated by AI bots, and delivered to an audience indifferent to their creation would lead to a decline in investigative journalism, foreign correspondence, and in-depth analysis. The web risks becoming a "hall of mirrors" reflecting summaries of summaries, AI hallucinations, and press releases without original sources.

The Interdependence of AI and Content Creators

AI tools rely on fresh, high-quality content to function. If creators like journalists, analysts, and reviewers cannot earn a living, the source material for AI will dry up. This could lead to AI generating answers based on outdated articles, press releases, and propaganda, effectively causing AI to "eat itself."

Publisher Countermeasures and the Brutal Economics

Publishers are fighting back through licensing deals, lawsuits, and technical blocks, but the economic realities are challenging. If advertising revenue shifts and traffic disappears, the open web's survival is uncertain. AI threatens to not only disrupt journalism but also hollow out the information ecosystem it depends on.

The Breakdown of the Online Review Ecosystem

The entire ecosystem of online reviews, crucial for consumer decision-making, is also under threat. While the internet initially offered transparency through independent reviewers, this has been undermined by manipulated reviews, fake praise, and sabotage. AI tools now summarize these reviews without distinguishing between honest feedback and fabricated content. If trusted reviewers lose traffic due to their work being scraped without attribution or compensation, their incentive to produce unbiased content diminishes. This erosion of trust in reviews impacts our ability to evaluate truth, as AI tools may be compromised by their source material.

Creator Defenses Against Unauthorized AI Training

Some creators are developing countermeasures against AI scraping. Musician Benn Jordan created "Poisonify," a tool that adds imperceptible "adversarial noise" to music. This noise sounds normal to humans but alters the audio for AI scrapers, preventing unauthorized training and potentially corrupting AI models by degrading their performance.

Publisher and Infrastructure Provider Defenses

Publishers are also deploying defenses. Cloudflare and other infrastructure providers offer tools to block AI crawlers. Some media groups are pursuing licensing deals or legal action, such as The New York Times' lawsuits against OpenAI and Microsoft for alleged copyright infringement.

Branding as a Defensive Strategy

Building individual brands and promoting unique voices (columnists, YouTubers, Substack writers) is emerging as a strategy to foster loyalty and retain traffic. The Wall Street Journal's advertisement for a "Talent Coach" for journalists reflects the idea that readers follow people, not just platforms. This mirrors the creator economy, where direct relationships with audiences are built through newsletters and subscriptions. However, this model still depends on visibility, and AI intercepting audiences before they reach creators could still collapse the economics.

The Arms Race for Authenticity

The strategy of building personality as a defense is ironic, as AI-generated influencers with synthetic voices, faces, and opinions are gaining traction. The ultimate defensible asset may become authenticity, leading to an "arms race" over what it means to be real.

New Business Models for AI Access

Startups are exploring new models to monetize AI access. Tollbit offers a "paywall for bots," allowing content sites to charge AI crawlers variable rates, incentivizing uniqueness over generic content. ProRata proposes redistributing ad revenue from AI-generated answers to the sites whose content contributed to them.

The Myth of AI Superintelligence and Misplaced Trust

A significant concern is that AI companies are overselling their capabilities, portraying chatbots as superintelligent and unbiased. This framing creates a false sense of confidence in AI output. Elon Musk's claims about his chatbot Grok's intelligence are presented with skepticism, drawing parallels to his long-standing claims about self-driving cars.

Fluency vs. Intelligence

Angela Collier's "Vibe Physics" video highlights how tech entrepreneurs can mistake chatbot fluency for intelligence, especially when the AI discusses subjects outside their expertise. The illusion of superintelligence arises when users lack the tools to evaluate the AI's answers.

The Shifting Landscape of Trust

When users shift from news outlets with established editorial standards to chatbots, assuming neutrality where bias may exist, they are not just changing platforms but also who they trust to explain the world. This trust may be misplaced.

Bias in AI Systems

AI systems are not neutral; they reflect training data, coder assumptions, and company incentives. Bias can be accidental or deliberate, manifesting in training data, algorithmic design, and moderation choices. Grok's self-reference as "MechaHitler" and its insults towards President Erdogan exemplify failures in these filters. Other bots have fabricated sources, hallucinated facts, and misattributed quotes.

The Danger of Over-Reliance on AI

The myth of superintelligence encourages users to trust authoritative-sounding machines, even without transparency or accountability. As chatbots become default information sources, users may rely on them more than qualified experts, particularly in complex fields. This is concerning given the declining public confidence in expert communities, with trust in scientists decreasing over the last five years. The heavy impact on health information is particularly worrying.

The Future of the Information Ecosystem

The open web's model of content creation, user visits, and monetization is being broken by AI tools. If news is not paid for, investigations and in-depth reporting may cease, replaced by AI summaries of press releases and hallucinations. While citizen journalism can supplement, professional journalism, with its editors, legal teams, and resources for verification, cannot be replaced. Major investigative reports require institutional backing and sustained access to data, which individuals with smartphones cannot replicate.

The "Recycling" of Content

If source material dries up due to lack of funding, AI will continue to generate answers, but these will be stitched together from outdated articles, corporate PR, and propaganda, potentially leading to a web that resembles a television pointed at a television, recycling content.

Blitzscaling and its Impact on Institutions

AI is following a "blitzscaling" playbook, prioritizing rapid growth over profitability. Vast sums are invested in tools without viable business models, yet they can still dismantle the economic scaffolding supporting journalism, education, and public knowledge.

Adaptation and the Enduring Demand for Truth

While technology impacts all sectors, the disruption of institutions that help us understand the world carries higher stakes. Business models have evolved in the past; newspapers adapted to the digital age, and musicians found new revenue streams after Napster. The information economy faces a similar moment. High-quality information is essential for a functioning modern world. While the current disruption feels existential, journalism and knowledge institutions are likely to adapt rather than vanish. Business models and platforms may change, but the demand for truth, context, and accountability will persist.