Modern science is no longer used primarily to describe reality, but to construct models that are then treated as reality itself. This shift spans cosmology, climate science, public health narratives, and artificial intelligence.

Across these domains, models increasingly mediate how reality is interpreted, from theoretical structures in cosmology, through institutional policy models in climate science and public health policy, to algorithmic systems that shape perception itself.

In cosmology, the process begins with theoretical constructs introduced to resolve gaps between observation and existing models, gradually accumulating unobservable entities such as dark matter and dark energy in order to preserve theoretical coherence.

In climate science and public health, scenario modelling becomes embedded within policy frameworks, where speculative and catastrophic scenarios evolve into instruments of institutional planning and political decision-making.

In artificial intelligence, the process reaches its most complete form, where model outputs no longer merely assist interpretation but increasingly shape the very framework through which information is accessed and understood.

The result is a growing gap between direct observation and theoretical construction. 

This pattern does not arise in a social vacuum. Scientific models increasingly operate within institutional environments shaped by funding priorities, government agencies, corporate interests, media organisations, regulatory bodies, and technological platforms. Once a particular model becomes embedded within these structures, powerful incentives often emerge to preserve it. Careers, research funding, political agendas, commercial interests, institutional legitimacy, and prevailing narratives may all become tied to the continued authority of the framework. As a result, debates about scientific models often become intertwined with broader questions of technocratic control and institutional, political, and financial power.

The issue is not whether models are useful; it is whether modern science still recognises where models end and reality begins.

Cosmology provides perhaps the clearest example.

The Cosmology Problem: The Expanding Gap Between Observation and Theory

The dominant scientific narratives about the universe increasingly reveal more about the assumptions built into their models than about the universe itself.

Modern cosmology is often presented as one of the great triumphs of contemporary science. Its explanations increasingly depend on assumptions embedded within its models rather than direct observation.

The gap between observation and inference has widened. The gap between empirical data and theoretical scaffolding has widened to the point where explanation risks being replaced by paradigm preservation.

At the center of contemporary cosmology is the Big Bang framework. In its original form, the framework emerged from observations of redshift phenomena in distant astronomical objects. These observations were subsequently interpreted within the emerging cosmological framework as evidence for an expanding universe.

Over time, however, the framework has been repeatedly modified to address problems it cannot resolve on its own.

The Dark Universe and Invisible Assumptions

Here’s the key issue: modern cosmology now depends on entities that have never been directly observed and are known primarily because they are mathematically required for the dominant model to function.

Today’s dominant cosmological framework, commonly referred to as ΛCDM (Lambda–Cold Dark Matter), rests on a set of assumptions required to make the model work. These include ordinary matter, which constitutes only a small fraction of the universe; cold dark matter, a hypothetical unseen form of matter proposed to explain certain astronomical observations that do not fit easily within the standard model; and Λ, the cosmological constant—now interpreted as dark energy—invoked to account for certain astronomical observations that have been interpreted within the standard cosmological model as evidence for accelerating cosmic expansion.

Taken together, dark matter and dark energy account for roughly 95 percent of the universe as described by this model.

In other words, according to the prevailing model, almost everything in the universe is something we have never directly observed. Most of what the model says exists has not been directly detected as dark matter or dark energy in laboratory experiments.

A model in which 95 percent of the universe is invisible has achieved remarkable success in describing what cannot be seen. This model in which 95 percent of the universe is invisible raises obvious epistemic questions.

Dark matter and dark energy are inferred from discrepancies between theory and observation. Remove them, and much of the explanatory architecture of modern cosmology begins to unravel. In modern cosmology they are not optional hypotheses but structural necessities of the model. Mathematical coherence within a model, however, is not the same as ontological reality. A theory may be internally consistent while still leaving open the question of whether the entities it invokes actually exist.

Many readers assume that such additions represent steady scientific progress, but that assumption deserves scrutiny. When a theory survives only by multiplying unseen components, an important question arises: is the model explaining reality, or merely protecting itself?

Inflation, Singularities, and the Patchwork Model

A similar pattern appears in the treatment of singularities and infinities. The standard Big Bang model begins with a spacetime singularity—an infinity of density, temperature, and curvature. In physics, such infinities typically signal a failure of theory rather than a physical reality, and here they mark an acknowledged breakdown of known physics, where the equations cease to describe reality. 

Inflationary theory is often presented as a refinement or improvement of the Big Bang, but it does not replace it. Inflation proposes a brief period of extremely rapid expansion in the early universe to solve specific problems such as horizon uniformity and spatial flatness. To do so, inflationary models typically invoke inflaton fields—hypothetical scalar fields whose properties are adjusted to generate the required expansion. These fields are currently unobservable, highly model-dependent, remain hypothetical and have not been empirically confirmed.

But here’s the problem: inflation does not resolve the foundational questions; it relocates them. It presupposes an expanding universe, does not explain the ultimate origin of spacetime, and introduces additional speculative mechanisms to stabilize the framework.

Adding unseen entities to rescue a theory resembles bolting extra parts onto a machine that will not start—eventually one must ask whether the machine was ever the right design.

When Theories Survive by Adding Layers

At each stage, when observation and theory diverge, the response has been consistent: add new layers. When anomalies arise, new entities are proposed; when tensions persist, parameters are adjusted. The framework survives, but questions about its underlying assumptions remain, and understanding does not deepen.

Philosophers of science such as Thomas Kuhn and Imre Lakatos argued that scientific frameworks often survive by adding auxiliary hypotheses rather than resolving foundational problems; the growing layers of dark sectors and inflationary fields follow this familiar pattern. Lakatos described such frameworks as “research programs” that defend a hard core by adding a protective belt of auxiliary hypotheses.

Institutional Science and Model-Driven Incentives

Modern cosmology is also shaped by institutional and financial incentives that rarely receive scrutiny. Large-scale cosmological research depends heavily on government space agencies and centralized scientific institutions whose priorities favor ambitious, data-intensive projects organized around overarching theoretical frameworks, as reflected in ESA and NASA mission proposals.

Career advancement, grant allocation, and institutional prestige often become associated with prevailing theoretical frameworks, creating structural incentives that can discourage challenges to foundational assumptions, even when anomalies persist. 

This pattern is not unique to cosmology: similar dynamics can be observed in climate modeling and artificial intelligence research, where complex models, opaque assumptions, and institutional momentum can outpace empirical verification and public understanding. In such environments, models risk becoming self-reinforcing systems of belief rather than provisional tools for inquiry. 

In some cases, explanation risks giving way to the accommodation of anomalies rather than their resolution.

A scientific theory is meant to illuminate reality, not merely absorb contradictions. When unobservable entities become indispensable to preserving a preferred worldview, the boundary between empirical inquiry and metaphysical commitment begins to blur. This is not an accusation of bad faith; it is an observation about how complex theoretical systems behave under strain. The institutional organization of science and research funding tends to reward results that fit the dominant theory and discourage results that challenge it, even in the absence of any deliberate deception.

Cosmology, Meaning, and the Question of Intelligibility

Contemporary cosmology implicitly adopts a form of reductionism in which matter and energy are treated as the ultimate explanatory ground. That conclusion isn’t something you can measure with a telescope—it’s a worldview assumption about what counts as a valid explanation.

Long before modern physics, classical philosophical traditions treated cosmology not merely as a technical problem, but as a metaphysical one. For example, the ancient Sanskrit text Bhāgavata Purāṇa describes the universe as an ordered hierarchy governed by intelligible principles, where physical structure, law, consciousness, and purpose are integrated rather than isolated. 

The significance of such perspectives is their philosophical orientation: order and intelligibility are taken as fundamental, not accidental. Physical laws are expressions of deeper structure, not brute facts that emerge from chaos. 

Modern cosmology, by contrast, is often interpreted within a materialist framework in which matter and energy are assumed to give rise to law, coherence, and consciousness through fundamentally unguided processes—a proposition that is philosophical rather than directly empirical.

Converging Intuitions of Cosmic Order

This intuition is shared by classical Western theology, which likewise held that cosmic order points beyond matter to rational grounding. The Gospel of John famously begins by identifying the Logos—often translated as Word or Reason—as the foundation of all things: “In the beginning was the Word… and all things were made through Him.” Similarly, the Apostle Paul writes that “in Him all things hold together” (Colossians 1:17). And in the Wisdom of Solomon 11:20 (Order by Measure and Number) (Deuterocanonical, and read by ancient Christian philosophers): “You have ordered all things by measure and number and weight.”

In both Eastern and Western traditions, intelligibility is not an accident of matter, but a reflection of deeper rational structure. In both Christian and Vedāntic traditions, cosmic order is grounded in an underlying rational or conscious principle, a view that aligns closely with classical metaphysical accounts of intelligibility and being. The shared assumption across these traditions is that reason precedes matter, rather than emerging from it—a philosophical position that contrasts with modern reductionist materialism.

The philosophical question is whether modern cosmology, by contrast, can explain why such intelligibility exists at all.

The Limits of Explanation and the Need for Humility

Here’s why this matters. The universe is not merely large or old; it is governed by stable laws, finely balanced constants, and mathematical relationships that permit complexity, life, and consciousness. These features are not incidental. They are the very conditions that make cosmology—and science itself—possible.

Yet within a strictly materialist framework, such order and intelligibility are treated as accidental outcomes of blind processes. We are asked to believe that chance produces not only matter and energy, but laws, coherence, beauty, and minds capable of discovering those laws. This is not an empirical conclusion derived from observation; it is a philosophical assumption embedded in the framework.

None of this negates the value of cosmological research. Measurement, modeling, and observation have yielded genuine insights. The problem arises when models are mistaken for reality and when provisional constructs harden into unquestioned truth.

A more disciplined cosmology would distinguish clearly between what is observed and what is inferred, between mathematical convenience and ontological commitment. Such restraint would not weaken science. It would strengthen it.

Science advances not by claiming to explain everything, but by knowing precisely what it can and cannot explain—and by resisting the temptation to mistake elegant models for ultimate reality. When cosmology recovers that humility, it may once again illuminate the universe rather than obscure it behind layers of abstraction. When models substitute for observation, science risks drifting from the study of reality toward the defense of its own abstractions.

Many of the cosmological issues discussed here—including dark matter, dark energy, inflation, singularities, and the growing gap between observation and theoretical necessity—are examined in greater detail in my recent book When Models Replace Reality: The Hidden Assumptions of Modern Cosmology.

Similar dynamics also appear in climate modelling, pandemic response, and artificial intelligence, where complex models, opaque assumptions, and institutional momentum can outpace empirical verification and public understanding.

Beyond Cosmology: Implausible Climate Modelling and a Broader Scientific Pattern

Climate science provides a particularly important example.

For more than a decade, one emissions scenario—RCP 8.5 and its successor SSP5-8.5—served as the foundation for thousands of climate-impact studies, government reports, policy initiatives, and media narratives. Governments declared climate emergencies. Pension funds and corporations reorganised investment strategies. Schoolchildren were repeatedly told they faced an existential threat. It influenced climate litigation, net-zero policies, ESG investment frameworks, adaptation planning, and educational materials.

In the official CMIP7 ScenarioMIP design paper published in 2026, Detlef van Vuuren and more than forty climate-scenario researchers wrote:

“For the 21st century, this range will be smaller than assessed before: on the high-end of the range, the CMIP6 high emission levels (quantified by SSP5-8.5) have become implausible…”

The significance of that statement is substantial. The shift became explicit during development of the CMIP7 framework that will inform the IPCC’s Seventh Assessment Report. Researchers responsible for designing the next generation of climate scenarios concluded that the highest-emissions pathway widely used throughout previous assessments no longer represented a plausible picture of the world’s likely future.

This matters because projections derived from that pathway influenced thousands of studies, climate-risk assessments, policy discussions, and public narratives. The issue is not that climate models exist. The issue is how speculative scenarios can become treated as de facto forecasts in policy and public discourse.

A scenario that helped justify climate emergencies and world-wide net-zero legislation is now described by the scientists designing the next generation of climate scenarios as implausible.

The significance extends beyond climate science. It illustrates how a model can evolve from a technical scenario into a framework that shapes institutions, public perception, and public policy on a global scale.

The issue is not that climate modelling is worthless. The issue is when assumptions embedded within models acquire authority that exceeds the evidence supporting them.

Readers interested in a more detailed examination of climate modelling, emissions scenarios, and the institutional dynamics may find these issues explored further in my book Climate CO2 Hoax.

Virology and the Pandemic Modelling Era

Much of the public justification for unprecedented lockdown policies rested upon epidemiological models projecting catastrophic outcomes. In Britain, projections associated with Imperial College helped shape government policy and were widely cited as justification for emergency restrictions. Similar modelling exercises influenced decision-making across much of the Western world. 

Mathematical models increasingly acquired authority over direct observation. The catastrophic projections that underpinned unprecedented public-health interventions increasingly came into question as many of the assumptions underlying the official narrative were themselves disputed.

Regardless of one’s interpretation of the events of 2020 and their underlying causes, the episode demonstrated how speculative models can rapidly acquire extraordinary political authority. Governments restricted movement, closed businesses, and suspended long-standing liberties largely on the basis of modelling projections rather than evidence.

The issue extends beyond any particular public health narrative. It reflects a broader shift in which models increasingly mediate policy decisions.

In such contexts, models do not merely inform policy; they begin to shape the interpretive frameworks through which reality is understood.

This article does not attempt to examine the broader controversies surrounding modern virology or competing interpretations of the events of 2020. Those questions fall outside its scope and are examined separately in my book No Worries, No Virus.

Artificial Intelligence and Synthetic Knowledge

AI is frequently presented as a neutral source of information. In reality, every AI system is trained on selected data, shaped by human decisions, and constrained by institutional priorities. It reflects assumptions about what information matters, which sources are authoritative, and what conclusions are considered acceptable.

Increasingly, people obtain information not through direct investigation but through AI-generated summaries and interpretations. In this way, artificial intelligence increasingly mediates knowledge itself.

Just as cosmological models mediate our understanding of the universe, and climate models mediate our understanding of future risk, AI increasingly mediates our understanding of reality itself.

Large language models are not trained to discover truth. They are trained to reproduce patterns found within vast datasets. Their outputs therefore reflect prevailing assumptions, dominant narratives, and the boundaries of their training. Consensus increasingly acquires the appearance of objective knowledge.

What social media companies once enforced through armies of moderators can increasingly be implemented automatically through algorithms. Questions of history, science, economics, and public policy become filtered through systems that appear objective while concealing the assumptions embedded within them.

The danger is not occasional mistakes, but the gradual replacement of independent inquiry with machine-generated interpretation. Over time, people begin to consume interpretations and assumptions rather than evidence, summaries rather than investigation, consensus and conclusions rather than understanding.

A society that increasingly delegates judgment to algorithms risks losing the habit of judgment itself. Consensus reality is then shaped less by direct observation and public debate than by the outputs of increasingly complex systems few people fully understand.

These questions concerning artificial intelligence, algorithmic mediation, and synthetic knowledge are explored in greater detail in my books The AI Illusion and Staying Human in the Age of AI.

Conclusion: Where the Models End and Reality Begins

The issue is not whether models should exist. Science could not function without them. The issue is whether modern societies still remember the distinction between a model and the reality it seeks to describe.

As complexity increases, abstraction begins to replace direct engagement with reality, and models begin to acquire the authority of truth. 

The question is not whether our models are useful. The question is whether we still possess the intellectual humility to recognise where the models end and reality begins—and whether we retain the willingness to look beyond them.

*

Click the share button below to email/forward this article. Follow us on Instagram and X and subscribe to our Telegram Channel. Feel free to repost Global Research articles with proper attribution.

Mark Keenan is a former United Nations technical expert and an independent writer on science, technology, political economy, and human freedom. He is the author of When Models Replace Reality, Climate CO2 Hoax, No Worries No Virus, and the AI-related books The AI Illusion and Staying Human in the Age of AI. His books and articles are available at Reality Books and on Substack at markgerardkeenan.substack.com.

He is a regular contributor to Global Research.

Featured image is from the author

Global Research is a reader-funded media. We do not accept any funding from corporations or governments. Help us stay afloat. Click the image below to make a one-time or recurring donation.