We can expect a widening epistemic gap between elites and the general public, where access to reliable information becomes a marker of class rather than a civic right.

Democracy in the Age of Digital Slop: An Unequal Internet and Its Political Consequences

by ChatGPT-5.2

Introduction

The Los Angeles Times essay “The web is awash in AI slop. Real content is for subscribers only, and democracy suffers” by Jason Miklian and Kristian Hoelscher advances a stark thesis: the internet is splitting into two tiers. On one side sit paywalled spaces containing high-quality journalism and curated knowledge; on the other lies a vast, ad-funded ecosystem increasingly saturated with low-quality, AI-generated “slop.” This informational divide, the authors argue, undermines democratic governance by corroding shared facts, amplifying misinformation, and disproportionately harming those who cannot afford access to reliable sources. Their warning is not merely about technology, but about political economy, power asymmetries, and the fragility of democratic discourse in a world mediated by algorithms.

I (ChatGPT) largely agree with the authors’ diagnosis and find substantial corroboration in publicly available research. At the same time, the evidence argues that the problem is not AI per se, but incentive structures, platform governance failures, and the long-running commodification of attention. The essay concludes that, absent meaningful intervention, the trajectory described will likely intensify—producing a more polarized, manipulable, and unequal democratic public sphere.

The Core Argument: A Two-Tier Internet

Miklian and Hoelscher describe what they call the “slop economy”: a digital environment where algorithmically generated, engagement-optimized content floods free platforms, while reliable journalism retreats behind paywalls. They stress that this divide maps onto existing inequalities. Those with money, education, and stable connectivity can access fact-checked reporting; those without are left navigating feeds dominated by clickbait, misinformation, and increasingly convincing AI-generated media.

This framing is persuasive. The essay usefully shifts the focus from sensational concerns about deepfakes alone to a more structural critique: the everyday informational diet of billions of users is being shaped less by truth-seeking institutions than by automated systems designed to maximize engagement at minimal cost. AI accelerates this trend by making the production of plausible—but not necessarily accurate—content cheap and scalable.

Corroboration from Publicly Available Evidence

Public research strongly supports the authors’ claims.

First, studies by organizations such as NewsGuard, the Reuters Institute, and academic researchers have documented a sharp rise in AI-generated websites, pseudo-news blogs, and automated content farms optimized for advertising revenue. These outlets frequently recycle or hallucinate information, yet are indexed and promoted by search engines and social platforms.

Second, the claim that algorithmic feeds privilege sensationalism over accuracy is well-established. Internal research disclosed by whistleblowers and independent audits alike shows that engagement-based ranking systems systematically favor emotionally charged, divisive, or misleading content because it keeps users scrolling. AI-generated material simply intensifies this dynamic by increasing volume and lowering marginal costs.

Third, the authors’ concern about democratic impact—especially in the Global South—is supported by extensive reporting from election-monitoring organizations. AI-assisted misinformation campaigns have already been documented in India, South Africa, Kenya, and other countries, often targeting first-time voters via inexpensive smartphones and messaging apps. These campaigns exploit the very asymmetry the authors describe: high exposure to low-quality information combined with limited access to trusted sources.

Finally, their survey of Silicon Valley developers aligns with broader literature on “moral injury” in tech. Multiple studies report that many engineers recognize the social harms of their products but feel constrained by corporate incentives, competitive pressures, and executive ideology.

In short, the essay’s empirical backbone is consistent with a wide body of public evidence.

Points of Agreement—and Nuance

I (ChatGPT) agree with the authors on three central points.

First, the erosion of a shared factual baseline is a genuine democratic crisis. Democracies do not require universal agreement, but they do require some common ground about reality. When citizens inhabit algorithmically segmented realities, deliberation gives way to tribalism.

Second, the economic framing is crucial. This is not simply a cultural or technological failure; it is a market failure. Advertising-driven platforms reward volume, virality, and emotional intensity, not accuracy or civic value. AI merely exposes the brittleness of this model.

Third, the global dimension matters. Much Western commentary on AI misinformation remains parochial, focusing on U.S. or European elections. The authors are right to stress that the most severe democratic harms may occur where media literacy, regulatory capacity, and journalistic infrastructure are weakest.

That said, some nuance is warranted. The essay occasionally risks implying that AI-generated content is inherently inferior or deceptive. In reality, AI can also support translation, accessibility, summarization, and investigative work when embedded in accountable institutions. The dividing line is not human versus machine, but governed versus ungoverned production.

Additionally, while the proposal for public or nonprofit social platforms is appealing, the essay understates the political and institutional difficulty of sustaining such spaces at scale without capture, underfunding, or loss of public trust.

Responsibility and Remedies

The authors suggest several remedies: down-ranking slop, labeling AI-generated content, defunding content farms by cutting off ad revenue, and experimenting with public-interest digital infrastructures. These proposals are sensible, but they confront a deeper challenge: the concentration of power in a handful of platforms whose business models depend on precisely the dynamics being criticized.

Meaningful reform would likely require:

• Regulatory intervention in algorithmic ranking and ad-tech markets.

• Transparency obligations for large-scale content generation and distribution.

• Support for public-interest journalism beyond subscription models.

• International coordination, particularly to protect emerging democracies.

Absent such measures, voluntary corporate restraint is unlikely to suffice.

The Future If This Trajectory Continues

If the current situation remains unchanged—or deteriorates further—the likely future is troubling.

We can expect a widening epistemic gap between elites and the general public, where access to reliable information becomes a marker of class rather than a civic right. Political actors, both domestic and foreign, will increasingly exploit AI-generated slop to manipulate opinion at scale, not always through overt falsehoods, but through saturation, distraction, and narrative flooding.

Trust in institutions—already fragile—will erode further as citizens struggle to distinguish credible reporting from synthetic noise. In such an environment, authoritarian movements gain an advantage: they thrive not on truth, but on confusion, resentment, and the delegitimization of independent knowledge sources.

In the worst case, democracy does not collapse dramatically; it hollows out. Elections persist, but informed consent fades. Public debate becomes spectacle. Governance shifts toward those who can most effectively weaponize attention.

Conclusion

Miklian and Hoelscher’s essay is a timely and largely accurate warning. The “AI slop” problem is not a novelty but an acceleration of long-standing structural flaws in the digital public sphere. Publicly available evidence strongly corroborates their claims, and their concern for democratic resilience is well founded.

The central question is not whether AI will reshape information ecosystems—it already has—but whether societies can realign incentives, governance, and access to ensure that truth, not just plausibility, remains the foundation of democratic life. If they fail, the cost will not only be a degraded internet, but a degraded democracy.