AI Slop and the Tragedy of the Commons

VZ Lens

Through a VZ lens, this analysis is not content volume - it is operating intelligence for leaders. As low-quality content floods ecosystems, shared information trust decays. The strategic response is quality governance, not publishing volume. The practical edge comes from turning this into repeatable decision rhythms.

TL;DR

Garrett Hardin’s “tragedy of the commons” model from 1968: if everyone is free to use a shared resource, everyone will overuse it, and in the end, no one will be better off. The internet’s content ecosystem is currently experiencing this. AI is the catalyst for overgrazing—and the commons is being depleted. This dynamic is not merely a technological or content-related problem, but a fundamental flaw in system design, where individual incentives cause collective harm. This article examines how this mechanism works, drawing on historical analogies, economic theory, and the specific characteristics of the digital ecosystem.

---

A Pasture and a Search Engine

The crypt of the Esztergom Basilica, silence. I go up and look at the Danube from the lookout. On the other bank: Párkány—Štúrovo. The border between two worlds. The physical world is clear here: one side is Hungary, the other is Slovakia. But in the digital world, there are no borders—and that is the problem.

Garrett Hardin wrote about the tragedy of the commons in 1968. Anyone can graze on the commons. Every shepherd makes a rational decision: if I add one more cow, my profit increases. But if everyone thinks this way, the pasture becomes depleted. Individual rationality leads to collective catastrophe.

The internet’s content ecosystem was a common pasture. Anyone could publish freely. That was the promise of democratization. But AI has changed the arithmetic: one person, with one AI, can generate thousands of pieces of “content” daily. The common pasture can’t handle it.

What is the tragedy of the commons, and why is it a flawed fate?

Before we apply this to the internet, we must thoroughly understand Hardin’s model, which was originally quite deliberately simplified. Hardin’s story is about a village where every villager is free to use the common pasture with their own cows. It is a rational decision for every villager to bring one more cow out to the pasture: the extra benefit is theirs alone, while the burden (the additional grazing of the pasture) is shared by the community as a whole.

[CORPUS] — Unknown: “Picture a village arranged in a preindustrial style around a large, central, shared plot of land called a commons. The villagers use this land mostly for grazing sheep and cattle, which they subsequently shear, milk, or slaughter for their own sustenance or profit. Because the commons isn’t owned…”

This way of thinking, however, inevitably leads to disaster. As Hardin writes: “the tragedy of the commons has become a symbol of environmental degradation, where numerous individuals share a scarce resource.” [CORPUS] The tragedy is not a matter of chance, but a logical consequence of the system’s structure. Everyone acts rationally from their own perspective, but since the system cannot internalize external costs (other people’s cows receiving less feed), the aggregate of all decisions produces an irrational outcome. This is an external effect not addressed by classical market mechanisms.

The model is critically important because we see similar dynamics in the real world: overfishing, deforestation, and air pollution are all variations on the tragedy of the commons. Hardin and others, such as H. Scott Gordon in his study on fishing, have pointed out that free access to resources often leads to their devaluation.

[CORPUS] — Unknown: “The fish in the sea are valueless to the fisherman, because there is no assurance that they will be there for him tomorrow if they are left behind today. (Gordon 1954, p. 124)”

For a long time, internet content did not appear to be a scarce resource. Server capacity and bandwidth are expanding, but attention is finite. The commons metaphor becomes more accurate when we consider users’ finite attention and platforms’ finite discoverability as the truly scarce resources.

How did the internet become a content commons?

The early promise of the internet was the democratization of knowledge: anyone could publish, anyone could access. This created a vast, virtual commons where the “grass”—valuable information, personal experiences, professional analyses—could grow freely. Blogs, forums, and later social media platforms functioned as a kind of self-regulating ecosystem. Quality content naturally rose to the top (through links and shares), while the noise remained at the bottom.

Search engines like Google took on the role of digital gatekeepers. Their algorithms—based on the PageRank principle—sought to place the best, most relevant content at the top. This can also be viewed as a form of privatized regulation: the platform (the search engine) established the rules for access to the commons. For many years, this worked quite well.

AI slop—cheap, industrially generated, worthless content—has radically altered this balance. In Hardin’s model, this corresponds to the technological leap whereby a shepherd could suddenly drive not one, but a thousand cows onto the pasture, without incurring a significant portion of the costs. The scale of individual rationality also changes: for a content creator, AI enables the production of a much larger volume of content at a much lower cost (in terms of time and expertise) than before, in the hope that the platform’s algorithm will discover and monetize it. This is the rational decision.

But when this decision is made not by one person, but by hundreds of thousands or millions, the commons—the digital space for attention and discoverability—quickly becomes exhausted. The grass (valuable content) disappears beneath the noise (AI slop).

What economic principles are at work in the background?

The tragedy of the commons is a special case of the public goods problem. Public goods have two characteristics: they are non-excludable (it is difficult to prevent someone from using them) and non-rivalrous (one user’s consumption does not reduce the amount available to others—at least up to a point). Clean air or outer space are examples of this.

However, internet content is a shared resource that is rivalrous: if a user occupies the top spot on a search results page with an AI-generated article, they displace another, potentially high-quality article from that spot. Attention is finite. That is why the model is more accurate.

Toby Ord draws a parallel between this type of collective action problem and existential risks in The Precipice draws a parallel between this type of collective action problem and existential risks:

[CORPUS] — Unknown: “Consider the tension between the prisoner’s dilemma and the tragedy of the commons, where each individual’s incentives push them toward a collectively disastrous outcome.”

We are not talking about the extinction of humanity here, but the mechanism is the same: we are talking about a social trap. A social trap is a situation where individuals’ short-term self-interest conflicts with the group’s long-term collective interest. The fisherman who doesn’t catch everything he can today fears that someone else will catch it all tomorrow—so he catches it all today. This is a self-perpetuating, destructive cycle.

[CORPUS] — Unknown: “A social trap is a conflict over what is the best use of any given resource for the interest of the individual, as opposed to the common or collective good. It was once in the best economic interests of fishermen, herdsmen, farmers… to catch all the fish they could… Yet when individuals act independently to maximize profit, they u” [UNVERIFIED]

In the case of AI-generated content, the short-term individual interest is as follows: generate as much content as possible (even of minimal value), secure clicks, ad revenue, and algorithmic visibility. The long-term collective interest, however, would be for the internet to remain a reliable medium providing valuable information. The former is stronger because the reward is immediate and individual, while the harm is delayed and distributed among everyone.

Why are search engines unable to stop over-optimization?

Hardin proposed two classic solutions to avoid the tragedy: regulation (e.g., quotas, central decision-making) or privatization (allocating ownership of the commons, so the owner has an interest in maintaining it). On the internet, search engines have tried to function as privatized gatekeepers. Google privatized access regulation through its own algorithms.

The fundamental weakness of this system is now becoming apparent. Search engine algorithms rely on external, formal indicators: keywords, link structure, page speed, user behavior. These signals originally indicated the quality and relevance of the content. However, AI spam is capable of perfectly mimicking these formal signals without offering any underlying value or original thought.

It’s as if every cow in the pasture looked perfect and made the right sounds, but in reality produced neither milk nor meat—only consumed the grass. The shepherd (the search engine) cannot distinguish the valuable animal from the worthless one based on external signs.

Nassim Taleb describes a similar, though not entirely identical, phenomenon in Fooled by Randomness:

[CORPUS] — Unknown: “The opportunity cost of missing out on the discovery of a truly new thing—like the airplane or the car—is negligible compared to the poisoning you get from sifting through all the junk to get to these gems.”

AI slop causes precisely this “poisoning”: it poisons the information-searching process. The user or search engine must sift through an ever-increasing amount of content clutter in order to find fewer and fewer gems (valuable content). The search engine’s failure as a gatekeeper lies in the fact that its current tools are incapable of measuring the intrinsic value of content, the depth of knowledge, or the authenticity of the author’s intent. They see only the external form, which AI can perfectly mimic.

Are there successful alternatives to avoid the “tragedy”?

Hardin’s model dominated thinking for a long time, but international scholars such as Elinor Ostrom (who received the Nobel Prize in Economics for this work) have demonstrated through empirical research that the tragedy is not inevitable. In the real world, many communities have successfully managed their common resources without the need for central regulation or complete privatization.

[CORPUS] — Unknown: “I then pose theoretical and empirical alternatives to these models to begin to illustrate the diversity of solutions that go beyond states and markets. Using an institutional mode of analysis, I then attempt to explain how communities of individuals fashion different ways of governing the commons.”

Ostrom identified eight design principles for successful commons management, including: clearly defined boundaries, rules appropriate to local conditions, collective decision-making, community oversight, graduated sanctions, conflict resolution mechanisms, and recognition of self-regulation by external authorities.

These principles also provide a framework for saving the internet’s content commons. The role of “gatekeeper” need not necessarily fall to a central algorithm, but rather to decentralized, community-based filters. Examples exist: Stack Overflow (before it was flooded with low-quality questions) operated precisely with such a community-moderated reputation system. Wikipedia does as well, though it too struggles with AI-generated content. These are digital village communities that have established their own rules to protect their shared resource (the platform’s quality content).

The challenge is that this community-based approach does not fit the scale and profit-driven nature of most commercial platforms (Google, Facebook, YouTube). Their algorithms seek to maximize (attention, time spent on the platform), not quality or community well-being.

What new filters do we need in the age of AI slop?

If search engines, acting as technocratic gatekeepers, have failed, and central regulation (e.g., a general ban on AI content by platforms) is practically unfeasible, then the solution must emerge on the user side. Users must develop new “senses,” new filters, to evaluate content.

1. Author Reputation, Not Anonymity: In the future, the value of content will increasingly be determined by the credibility of the source. A well-written article is not enough. Who is behind it? What is the author’s background and professional experience? Is there a traceable digital footprint that demonstrates consistency and accountability? In the online world, this is a form of digital shepherding: those with a long-term stake in the health of the pasture take better care of it.

2. Community Validation, Non-Algorithmic Promotion: The value of a piece of content is indicated not by its visibility (number of shares), but by the quality of discourse it generates. Who is referencing it? In what circles (professional communities, closed forums) is it spreading? Twitter’s old “intermediary” layer or professional Subreddits can play such a validating role, but new platforms specialized for this purpose may also be necessary.

3. Research Depth and Originality, Not Formal Optimization: Content must demonstrate a chain of reasoning. Does it use sources? Is it aware of the counterarguments to the topic? Does it present new data or personal insights, or does it merely recycle others’ ideas? AI is currently very weak at deep, context-dependent analysis and truly original thought generation. This layer will be what distinguishes human-generated content from machine-generated content.

4. Restricted-Access Commons: A promising future trend is the rise of smaller, moderated platforms based on subscriptions or community membership. These are enclosed pastures where the community sets the rules, and the cost or commitment of membership filters out content creators driven solely by quantitative profit. Substack, professional Discord servers, or certain instances of Mastodon point in this direction.

For the average user, the most important filter will be simple: trust in the source. Just as we wouldn’t eat food bought from an unknown street vendor in the physical world, we won’t accept information from every anonymous or opaque source in the digital world. Rebuilding trust is the biggest challenge.

Key Takeaways

• Hardin’s “tragedy of the commons” perfectly models the dynamics of the AI slop: individual rationality (cheap AI content = lots of clicks) leads to collective catastrophe (the depletion of the content ecosystem).
• The early, democratic commons of the internet is being rendered unsustainable by the technological leap in mass content production (AI), much like historical overfishing or deforestation.
• Search engines, as privatized gatekeepers, have fundamentally failed: their algorithms measure external, easily manipulable signals of content, not its intrinsic value or originality.
• This tragedy is not inevitable. Successful community-driven models exist in the real world (Ostrom’s principles) that can provide guidance for the digital realm as well.
• The solution will not be a single technical fix, but rather systemic change: we need new filtering mechanisms (author reputation, community validation, depth of research) and possibly a shift toward new, smaller, curated digital communities.

Frequently Asked Questions

What is the tragedy of the commons in the context of AI slops?

Garrett Hardin’s 1968 commons model: if the pasture is open to everyone, everyone grazes as much as they can, and the pasture is ruined. The internet is a content commons: individual rationality (cheap AI-generated content = lots of clicks) leads to a collective catastrophe (content becomes devalued, quality is squeezed out). AI enables “overgrazing,” that is, the production of vastly greater quantities of content at minimal cost, which accelerates the depletion of the common pasture.

How does the AI slop affect content markets?

Search engines fail as gatekeepers: technical signals (SEO, formatting, links) can be manipulated. 21% of YouTube recommendations are already AI-generated. The result: users cannot distinguish genuine expert content from AI-generated copies, and as platforms’ algorithms adapt, bad content drives out the good. This reduces user trust, increases the cost of information search (time, energy), and in the long run may lead to users abandoning platforms or migrating toward more closed communities.

Is the tragedy of the commons inevitable? Are there no counterexamples?

It is not inevitable. Elinor Ostrom’s Nobel Prize-winning work presented numerous real-world examples where communities successfully and sustainably managed their shared resources without central authority taking over. Clear rules established by the local community, oversight, graduated sanctions, and conflict resolution are key. In the digital world, the early years of Stack Overflow or Wikipedia’s moderation system are similar, though imperfect, examples of this. The challenge lies in scaling these models to the entire internet.

What could be the solution to the flood of AI-generated content?

There is no single magic solution; rather, it requires the cooperation of multiple layers of a system:

1. Technical/Platform Level: Search engines and platforms must develop new metrics that measure content originality, author credibility, and the quality of community validation, not just quantitative indicators.
2. Community level: We must strengthen digital communities that, with their own rules and moderation culture, are capable of fostering quality discourse (e.g., professional forums, subscription newsletters, moderated social media feeds).
3. Individual/Consumer Level: The most important filter must be the user’s awareness and critical thinking. When consuming content, one must examine the source’s consistency, the author’s verifiable background, and the depth of the content’s research. Rebuilding trust is key.

---

Related Thoughts

• AI Slop: The Market for Lemons Returns (Akerlof)

---

Zoltán Varga - LinkedIn Neural • Knowledge Systems Architect | Enterprise RAG architect PKM • AI Ecosystems | Neural Awareness • Consciousness & Leadership The commons grazed, the AI slop feasted.

Strategic Synthesis

• Identify which current workflow this insight should upgrade first.
• Set a lightweight review loop to detect drift early.
• Review results after one cycle and tighten the next decision sequence.