Friends,
I urge you to read this deeply concerning and important essay by Dr. Paul N. Edwards, Professor Emeritus of Information, History, and Science and Technology Studies:
Edwards, P. N. (2026). The end of science? Agnosis and American fascism. Science, Technology, & Human Values. https://doi.org/10.1177/01622439261474980 [posted below]
A key term in the title is agnosis, which Edwards uses to describe the deliberate production of ignorance, nonknowledge, or antiknowledge. Just as important is his idea of knowledge infrastructures: the institutions, people, methods, archives, and professional norms that allow societies to produce reliable knowledge over time.
This is what makes the essay so relevant to higher education. Edwards asks us to see attacks on research funding, federal data, scientific expertise, universities, and academic freedom not as isolated developments, but as threats to the very infrastructure through which a democratic society comes to know itself.
That argument resonates strongly with something I have been emphasizing on this blog: academic freedom is part of the infrastructure of civil rights. If we weaken the institutions that produce knowledge about inequality, discrimination, public health, environmental harm, and governmental action, we also weaken our capacity to document injustice and demand accountability.
For good measure, I am cutting and pasting Science and Technology Studies courses offered currently at UT as I fear, based solely on Edwards' essay, that these courses could be at risk of getting cut largely because of their interdisciplinarity and criticality:
STS Coursework at UT Austin
UT Austin features a formal "STS" course designation within its academic catalogs, bridging the liberal arts, social sciences, and technical innovation. Notable options include: [1]
- STS 318: How We Shape Discoveries and How They Shape Us (examining the cultural and technical complexities of energy, biomedicine, and nanoscience). [1]
- STS 319: Information Technology and Social Life (analyzing how digital tech mediates work, home, and social structures). [1]
- STS 332: Nanotechnology and the Science Revolution. [1]
- STS 380: Proseminar: Current Issues in the Societal Impact of Science and Technology (a graduate-level overview examining science as a human enterprise). [1]
There may be others, but this is what a quick Google search turned up.
The essay is sobering, but Edwards does not end in despair. His call is for scholars and citizens to defend evidence, preserve knowledge, protect institutions, and act.
I couldn't agree more. And it's what we've been doing—and consistently so—here in Texas.
It is a long read, but a very important one.
-Angela Valenzuela
The End of Science? Agnosis and American Fascism by Paul N. Edwards, September 2026 issue,
Abstract
The Society for Social Studies of Science (4S) annually awards the John Desmond Bernal Prize to one or more individuals who have made distinguished contributions to the field of Science and Technology Studies. Past winners have included founders of the field, along with outstanding scholars who have devoted their careers to understanding the social dimensions of science and technology. This article is the revised text of the 2025 Bernal Lecture by Paul Edwards.
In January 2025, fascists led by Donald Trump took over the United States federal government for the second time. Following the Heritage Foundation blueprint Project 2025, these far-right authoritarians colluded with mega-corporations, billionaires, and white-nationalist Christian evangelicals to disrupt the government, nullify democratic processes, and execute Trump's every whim as “policy.” His regime has successfully (if not entirely) evaded Constitutional checks and balances; manipulated corporate affairs, the US economy, foreign allies, and the entire rules-based world order by means of tariffs, executive orders, extrajudicial killings, threats of force, and war; created a huge, militarized secret police force (Immigration and Customs Enforcement); brought false charges against political opponents to neutralize and punish them; and, resuscitated the 19th-century colonial Monroe Doctrine to license renewed US imperialism. The Trump autocracy has brought major American institutions to their knees, singling out federally funded scientific enterprises and research universities for ideological assault via budget cuts. Incredibly, even as I write, leaders at the highest levels of our government are discussing how to suspend the 2026 elections by declaring a trumped-up national emergency.
These actions and characteristics are hallmarks of fascism, “a system of government marked by centralization of authority under a dictator, a capitalist economy subject to stringent governmental controls, violent suppression of the opposition, and typically a policy of belligerent nationalism and racism” (American Heritage Dictionary, 5th edition). Like many of you, I never imagined the United States would come to this. Naïf that I was, I believed that democratic norms and institutional guardrails would hold. I imagined, even more naively, that science and reason would ultimately prevail. Even as I feared the worst, I secretly hoped that US science and democracy would help steer the world away from a future of dangerous climate change, deadly pandemics, out-of-control artificial intelligence (AI), and ever more destructive wars.
Dear reader, I could not have been more wrong. As I receive with great humility this highest of prizes from the Society for the Social Studies of Science (4S), I can only plead that its namesake J.D. Bernal suffered from the same delusion. In 1939 (of all years), Bernal wrote that “we have in the practice of science the prototype for all human common action. The task which the scientists have undertaken—the understanding and control of nature and of man himself [sic]—is merely the conscious expression of the task of human society….In science men have learned consciously to subordinate themselves to a common purpose” (Bernal 1939). How I wish it were so.
Both that delusion and my fear of its failure arrived early in my life and simultaneously. I spent my teenage years in Staunton, Virginia, a small town of 18,000 in the Shenandoah Valley. It was a differently mediated world then, a world of books, magazines, newspapers, and three television channels. No email, no Internet, no online anything—and not a lot for a young person to do. So, I read, and read, and read. On the stairway landing just outside my bedroom, my parents positioned a bookshelf containing encyclopedias for children and young adults (the Britannica had pride of place in our living room, but that was reserved for school projects). I sat on that landing many nights, paging through the twenty-volume Book of Knowledge and The World Book encyclopedia.
My other love was the science fiction section of the Staunton Public Library. I read nearly every book. Those stories told of technological wonders “indistinguishable from magic,” as Arthur C. Clarke put it in 1962, some of them becoming reality before my bright young eyes. My childhood was the Atomic Age, promising nuclear energy “too cheap to meter,” along with nuclear submarines. Nuclear airplanes! Nuclear refrigerators! And nuclear weapons, too dangerous to use, but too powerful to abandon. I watched the first moon landing on TV at a summer camp in upstate New York. My daydreams were rockets to the stars, but my nightmares were of mushroom clouds rising over distant cities as I desperately sought escape. Computers, robots, and AI abounded in those novels. Miracle machines that could save the world. Or take it over. Or destroy it.
My political coming of age happened around the first Earth Day in 1970. I read my copy of The Environmental Handbook (Bell 1970) until it fell apart. By eighth grade, I had organized a school “Ecology Club” and led my first protest march to a construction site near the school where workers were struggling to ignite downed trees with motor oil, covering the school area with thick smoke.
And there you have my entire career in a nutshell. I became a “professor of world disaster,” as a student affectionately (?) called me. I’ll take the title, thanks.
Nightfall
One particular sci-fi story marked me deeply: Isaac Asimov's “Nightfall,” written in 1941—a particularly ominous year on Earth. In that story, the planet Lagash has six suns. Bathed in continuous daylight, its present-day inhabitants have never known a dark sky. In the story, scientists have discovered evidence of nine previous advanced civilizations—each one destroyed by fire, in a recurrent cycle lasting some 2,000 years. The broadly popular Cult of religious millenarians believes that this occurs when the planet periodically passes through a gigantic cave, where terrifying things called “stars” appear and rain down cleansing fire.
The scientists have analyzed irregularities in the planet's orbit, revealing the existence of a previously undiscovered moon. And they have calculated that a total eclipse of all six suns at once is imminent—right on schedule after 2,049 years. The scientists explain their theory to the Cult, hoping to reassure them that the planetary dive into darkness will not last long and should not be feared. Instead, the Cult denounces their explanation as blasphemy because it undercuts the “divine will” of the stars. They see science as their enemy.
The eclipse begins. As the planet slowly plunges into darkness, the stars emerge. The populace goes mad with terror. A torch-bearing Cultist mob marches on the observatory. The scientists prepare to defend it. But they too go mad with terror. And the fire consumes it all, again.
In the United States, we have seen nightfall coming for more than a decade. I first wrote about it back in December 2016, following the first election of Donald Trump—ten years, a lifetime ago (Edwards 2016). Even before taking office, the then-incoming administration sent a seventy-four-point questionnaire to the US Department of Energy (DoE). At the time, DoE housed several major climate laboratories, hundreds of climate scientists, and the scientifically crucial Program on Climate Model Diagnosis and Intercomparison (Edwards 2010;, 2012). The questionnaire sought exhaustive background on every DoE scientist, including their unpaid positions, blogs, and websites to which they contributed. The new administration also demanded a list of DoE employees and contractors who participated in the Conferences of the Parties to the United Nations Framework Convention on Climate Change or in the Interagency Working Group on the Social Cost of Carbon, a government entity tasked with accounting for the overall costs and benefits of limiting anthropogenic climatic change.
Given the new president's well-known antagonism to the very idea of anthropogenic climate change, these were ominous signs. “A dark time is coming to American climate science,” I wrote then (Edwards 2016). “Trump's mob of climate change deniers has begun its march on our present-day observatories.” The following month, on his first day in office, Trump withdrew the nation from the landmark 2015 Paris Agreement.
Knowledge Infrastructures (and How to Destroy Them)
Before continuing my main theme, let me pause briefly to introduce this concept. In general, infrastructures provide basic, widely shared systems and services such as water, power, transport, telephone, and Internet. By extension, in the 2000s and 2010s, my research group1 defined knowledge infrastructures as “robust networks of people, artifacts, and institutions that generate, share, and maintain specific knowledge about the human and natural worlds” (Edwards et al. 2007; Edwards 2009; Edwards 2010, 7; Edwards, Jackson, and Chalmers 2013).
More than a vague analogy, every word of this definition is meaningful and carefully chosen. In A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming (Edwards 2010), I refined our definition to center routine production of reliable, widely shared, usable knowledge. These qualities distinguish knowledge-producing infrastructures from information management infrastructures such as libraries, platforms, and databases, as well as from research infrastructures that generate new findings still subject to controversy, requiring refinement, confirmation, and testing.
In my usage, true knowledge infrastructures exhibit considerable stability. Methods and findings change over time, but changes are incremental rather than radical. Preserving historical data and maintaining continuity through calibration and quality control is a major concern; practitioners are expected to be honest, apolitical, and meticulous. Knowledge infrastructures maintain archives of their own, ensuring that future users can reconstruct logics, correct errors, and continue to build on the existing base. Because they produce useful, trusted knowledge, many are effectively “public goods” that play important roles in public policy and receive government support.
Knowledge infrastructures I frequently use as examples include, for example, the Centers for Disease Control (CDC), which track outbreaks of infectious diseases based on data collected from doctors and hospitals, and predict the velocity and direction of epidemics. The National Weather Service forecasts weather throughout the United States and worldwide, using powerful computer models to analyze global data. Together with the National Aeronautics and Space Administration (NASA), the National Oceanic and Atmospheric Administration (NOAA) creates and manages satellites that provide global data on key climate indicators such as top-of-atmosphere energy balance and vertical temperature profiles. These models and data serve as input to the highly respected assessments of the global Intergovernmental Panel on Climate Change.
Other examples include the Bureau of Labor Statistics, the Food and Drug Administration, and the US Census Bureau. Knowledge infrastructures need not be government agencies. Some private-sector entities, such as investment analysis firms like Morningstar, also serve this function. Once again, the key is that they impartially generate nonpartisan, reliable, widely used knowledge on a routine basis using standardized methods and data.
Knowledge infrastructures can never be fully insulated from political influence, investigator bias, and other forms of corruption, but until recently the US governmental system safeguarded those it operated. Expert agencies such as those mentioned above were legally protected from political interference. As a result, before Trump, frustrated leaders who were hostile to these agencies’ knowledge claims sought to undermine their outputs. As documented by numerous historians, sociologists, and STS scholars, “merchants of doubt” (Oreskes and Conway 2010) in the pay of profitable industries disputed knowledge about the disastrous effects of leaded gasoline, tobacco smoke, ozone-depleting chemicals, acid rain, and climate change long past unequivocal scientific confirmation. The century-old strategy, often fully explicit in internal communications, involves exploiting uncertainty to generate public doubt and distrust, through advertising, lobbying, secrecy, and support of “contrarian” scientists (Oreskes and Conway 2010; Proctor 2012; Franta 2021a, 2021b). Lather, rinse, repeat. Taking a more direct approach, political appointees sometimes intervened to soften or delete language in agency reports, or to suppress them altogether.
In 2019, I argued that the first Trump administration's “innovation” was to interfere directly with inputs to climate science (Edwards 2019). For example, it proposed to eliminate five satellite programs from NASA's Earth science budget—all of them directly related to climate science. One of many glaring cases of decisions based entirely on political animus involved the Deep Space Climate Observatory satellite already in orbit. Like most satellites, it carries multiple instruments, such as one designed to provide early warning of solar magnetic storms. The administration proposed to defund data analysis for only the platform's two Earth-climate-related instruments. In other words, those instruments would continue to beam the data they collected back to Earth, but NASA would be prevented from analyzing them—all to save just USD3.2 million, mere pocket change in the grand scheme of national budgets. At the time, the Trump proposals for large budget reductions mostly failed to win the day. Congress simply ignored them and kept funding levels relatively stable.
Having learned from that failure, Trump's second administration began with savage attacks on major science institutions and elite universities using executive orders. Rather than wait for Congress to debate and approve his proposed federal budget, Trump and his minions immediately and unilaterally interrupted funding for major knowledge infrastructures, including all of those mentioned above, and for research science across the board. Elon Musk's short-lived tenure at the new, dubiously legal “Department of Government Efficiency” (DOGE) left a trail of devastation. DOGE illegally canceled leases on agency buildings and forcibly removed agencies to new facilities, reserving its heaviest blows for agencies researching the environment, public health, and climate change, such as NASA's Goddard Institute for Space Studies and NOAA's Geophysical Fluid Dynamics Laboratory, two crown jewels of modern climate science. DOGE employees zeroed out thousands of federal grants based entirely on a now-notorious keyword list and advice from ChatGPT (Custer 2026). By May 2025, the National Science Foundation (NSF) had terminated over USD1 billion in awards, without explanation. By the end of that year, over 10,000 STEM PhDs had been fired, retired, or resigned across fourteen research agencies. On a percentage basis, the NSF took the largest hit by far, losing 40 percent of its STEM PhDs (Mervis and Hersher 2026). The NSF budget approved by Congress in February 2026 restored most of its funding, but after the huge loss of staff and leadership, the number of grants NSF awarded in FY25 and FY26 to date remains well below previous years (Figure 1). The situation at the National Institutes of Health is similar.
The Trump administration extorted elite universities, including Harvard, Columbia, Johns Hopkins, and the University of Pennsylvania, with threats to withdraw hundreds of millions in Federal grants, or to prosecute them for supposed antisemitism. He appointed unqualified ideologues to leadership positions, such as Robert Kennedy Jr. (Department of Health and Human Services) and Jay Bhattacharya (National Institutes of Health and presently also interim director of the CDC). Following the April 2025 resignation of NSF Director Sethuraman Panchanathan, the agency drifted leaderless for nine months before James O’Neill was nominated. If confirmed, O’Neill will be the first nonscientist ever to lead the NSF. In April 2026, Trump's FY27 budget zeroed out funding for the NSF's Division of Social, Behavioral, and Economic Sciences, which has funded hundreds of STS researchers, including me. In May, Trump fired all twenty-eight members of the National Science Board that oversees the agency. Although the federal courts have since ruled many of these actions illegal, the Republican majority in Congress did nothing.
Need I continue? As STS scholars, dear readers, I’m sure that you are already aware of these atrocities and the vast number of other insults to good governance of research science by the most lawless, nepotistic, and corrupt administration in American history. If you’re not furious, you should be enraged.
Agnosis as a Contact Sport
My colleague, mentor, and friend, the late climate scientist Stephen Schneider, once wrote a book called Science as a Contact Sport (2010) about his life as a working scientist on the front lines of climate politics. Like him, we STSers must engage agnosis—the deliberate production of ignorance, nonknowledge, or antiknowledge (Proctor and Schiebinger 2008)—across a wide variety of fields. We’re not mere observers, spectators, or commentators here. We are players, and the game is rough; it can even be, without exaggeration, lethal.
When I started in STS in the early 1980s, the Strong Programme in sociology of knowledge dictated that scholars should bracket the truth of scientists’ claims. We were to ask not whether scientists’ claims are true, but how they are generated, spread, and sometimes overturned. By focusing on controversies that occur before consensus has formed, STS would shine a light on the process of knowledge production. So, we investigated practices of justification: not just “what is the evidence,” but what counts as evidence? When, why, and for whom? What role does evidence play in knowledge creation—as opposed to funding, prestige, groupthink, or the gender, race, personality, or cultural background of the scientists? The “symmetry principle” required us to ask after the social causes of both true and false beliefs.
As a method and epistemological commitment, the symmetry principle has served us well. Yet as astute STSers and commentators noted early on, in practice the symmetry principle always tended to uplift the minority view in a scientific controversy. If some said fluoridated water causes cancer, while many others said its dental benefits outweigh such a risk, an STS study could be understood as validating the minority view (Martin 1991; 1989). If some said vaccines caused autism, while many others found no link, following the symmetry principle might legitimate the antivaccine side. Whatever its scholarly merits, in politics the neutral observer stance is readily converted to validation of the losing side. It was largely for this reason that the so-called “third wave” of STS saw some of the field's most radical relativists change their tune, recognizing the need for unfettered, independent scientific expertise on matters of concern (Collins and Evans 2017; Latour 2018; Caudill 2020).
In today's deadly contest over the future of science itself, sitting on the sidelines is capitulation. No reader of this journal is a bystander. Like it or not, we are implicated, and we must take the field.
Are We Witnessing the End of Science?
Let's look at some of the forces of agnosis at work today. I’ll start with more about the Trump assault on science—but I warn you, this is not the whole story, and the story is not just about the United States.
Ideological Oversight and “Alignment”
Guided by the Federalist Society and Project 2025, the Heritage Foundation's blueprint for a fascist United States, Trump's pliant sycophants embraced the fringe legal theory of the “unitary executive.” That theory confers upon the president a king-like authority over the entire executive branch, including the expert agencies discussed earlier, no matter what Congress may have intended when it created them. The agenda laid out in Project 2025 envisaged replacing tens of thousands of career civil service employees—who may work for decades, serving Republican and Democratic administrations alike—with political appointees aligned with the party in power. While unable to fully implement that agenda due to pressure from federal employee unions and the courts (not to mention its own incompetence), the Trump administration's disruptions of expert agencies sent an unmistakable message: resist and you could lose your job. Ideological alignment was rapidly enforced through a series of Executive Orders (EO). “Improving Oversight of Federal Grantmaking” (EO 14332) requires these same political appointees to ensure that grants “demonstrably advance the President's policy priorities.”
Against a mountain of scientific evidence to the contrary, “Ending Radical and Wasteful Government DEI Programs and Preferencing” (EO 14151) declared that “the official policy of the United States government [is] that there are only two genders: male and female,” thus denying the very existence of intersex and transgender citizens. The same order led to the infamous list of prohibited “woke” terms, which included climate change, gender, inequality, and diversity (Yourish et al. 2025). EO 14303, “Restoring Gold Standard Science,” upholds a list of norms—including transparency, openness, reproducibility, and falsifiability—to which virtually all scientists would assent. But the same order also directs agency heads and their designees—political appointees—to actively “correct scientific information” and “discipline” scientists they find to have violated guidelines. Here, as everywhere, the real questions regard who decides whether principles have been followed, how they make that determination, and what consequences those people can mete out to scientists they decide have violated them. Under such conditions, is it at all surprising that 75 percent of scientists polled by Nature were considering leaving the United States? (Witze 2025).
Numerous existing scientific reports have been scrubbed from government websites. For example, the five Congressionally mandated US National Climate Assessments (NCA) conducted since 2000 no longer appear; the sixth NCA was already underway when the Trump administration dismissed its scientists and pulled its funding (Plumer and Dzombak 2025). In June 2025, the climate.gov website sponsored by NOAA was reorganized and replaced with www.noaa.gov/climate “in accordance with Executive Orders.”2 Many datasets and articles once presented on climate.gov are either no longer available or much harder to find. Perhaps even worse, many entities that once collected data have been ordered to cease doing so, in what ProPublica has called a “war on measurement” (MacGillis 2025). It is difficult to find a full accounting, in any field, of exactly what has happened to the data and reports managed by federal expert agencies.
We do not even have data about what has happened to our data. But the Trump assault on knowledge institutions is just one of many forces of agnosis currently undermining science. Here are a few of the others, equally if not even more destructive because broadly international.
Generative AI
First, a caveat: I think AI and machine learning are amazing. One day, they may do great things for humanity, if they do not destroy us first. Today, however, most public attention is fixed on the subset known as generative AI, especially large language models (LLMs), which are only one manifestation of a much larger set of AI techniques. My view of today's LLMs accords with that of the nameless computer scientist who called them “'sounds like an answer’ machines.” They “generate” not facts and truth but rather factoids and truthiness. They’re like lazy undergraduates: for them, fake citations, wrong calculations, and buggy code aren’t errors at all. So long as the products look like what a knowledgeable human might produce, they’ve succeeded.
Professional scientists and peer reviewers are using these tools now. So are some of you, though I am certain you check and recheck the results (right? You do check, don’t you?). In a recent study of over 22,000 scientific articles published from 2021 to 2025, between 7 and 22 percent of abstracts showed evidence of LLM use (Liang, Zhang, and Wu 2025). Maybe using AI to write an abstract of your paper seems like a good idea. It might even be a good idea. I know I hate writing abstracts, and you probably do too.
But still. Abstracts are just the thin edge of the wedge. Journalists and scholars have already detected thousands of examples of published science (and law, medicine, etc.) containing AI-generated fake citations—one of the few indicators where the accuracy of AI-generated material can be unequivocally verified at scale. Nature recently warned that these fake citations are propagating from one paper to others as AI-powered research assistants find and refer to them as if real, thus “polluting the scientific literature” (Naddaf and Quill 2026). How much of the body text, figures, and data in scientific publications is already AI-processed or AI-generated? How much of that was carefully checked and revised by expert human authors? How many peer reviews were written or assisted by AI? Preparing this for publication in mid-2026, it's already hard to know. And as AI improves, differentiating between its contribution and those of human authors will get even harder.
LLMs make many errors; in addition to citations, they are said to “hallucinate” nonexistent names, places, dates, events, and much more. Some LLM hallucinations, like impossible dates or wrong unit conversions, are relatively easy to spot, while others are more subtle. Where the subject is well known and audiences are large, crowds of human readers will often catch these errors. But in highly specialized fields, where only small numbers of people possess the knowledge required to recognize them, subtle mistakes can spread widely. In the traditional reputation economy of science, those mistakes will spread further and faster when presented as fact by well-known scholars. A perpetual question for the coming decades will thus be whether responses from the “sounds like an answer” machine are true, or just truthy—and how we can tell the difference.
Paper Mills, Fraud, and Misconduct
Measurement is fundamental to all science. But another contemporary form of agnosis stems, ironically, from measurement itself, in the form of “key performance indicators” and their ilk. Academics must prove they are “productive,” and the standards of science dictate an objective measure of productivity. Anything less, some say, invites favoritism. So, into our annual faculty reports go numbers of publications, journal impact factors, and indices based on citation counts, despite well-known, fundamental flaws in virtually all such measures (Brembs, Button, and Munafò 2013). Indicators like these—sometimes mandated at the state or even national levels—often matter more than anything else for promotion, tenure, and “merit” salary increases in many research universities. Yet as Goodhart's Law famously states, when a measure becomes a target, it immediately ceases to be a good measure, because (like “teaching to the test”) it changes behavior. People simply do whatever will improve their score. Audit culture rules. And as ever, evaluators far too often prefer simple but arbitrary measurement schemes to more complex but more accurate ones.
The mania for measuring productivity leads inexorably to widespread cheating. One salient example: the rise of for-profit “paper mills.” These entities are quite open about what they are doing. Not only can you pay to “author” a paper you didn’t write—you can choose which position you want in the author list. The so-called “paper” has already been written for you, perhaps even by an AI (COPE 2022). Such operations are phenomenally successful. A recent large study calculated that the rate of production by paper mills is growing ten times faster than the rate of legitimate publication. The published study's conclusion is well captured in its title: “The entities enabling scientific fraud at scale are large, resilient, and growing rapidly” (Richardson et al. 2025). One of its authors says we have reached “a do-or-die moment for the scientific enterprise” (Richardson 2025).
Scientific fraud has always existed, but in recent years it has exploded in tandem with the measurement mania. In a recent survey of 6,200 Chinese medical residents, over 46 percent self-reported engaging in at least one of the following fraudulent practices: buying or selling papers, plagiarizing, falsifying data, inappropriate attribution of authorship, or other misconduct (Chen et al. 2024). My point here is not to pick on Chinese medical residents, who are doubtless just as overwhelmed as junior scientists elsewhere. Instead, these numbers are just one of many indicators that basic norms are breaking under the pressure to publish.
What about the future scientists among our undergraduates? In my experience at Stanford University and the University of Michigan, cheating was already rampant long before ChatGPT. Now it is not only epidemic, but systemic, as institutions purchase AI subscriptions and encourage their use. Many professors take the easy path of urging students to do their own work while also encouraging them to learn “proper” use of AI tools. To promote genuine learning, some of my colleagues—no doubt like many of you, dear readers—have returned to handwritten blue books or oral exams, since no test or exercise done on a computer is immune to assistance from the “sounds like an answer” machine.
Science Education and Populist Pseudoscience
It's easy for educators like us to rally under the banner of “more education.” But under American fascism, what will coming generations of US students learn about science? For many of them, science education at the high school and college levels already includes slick “educational” videos from Prager “University.” Founded in 2009, Prager is a conservative organization with deep pockets, originally funded by fracking billionaires. Its videos are steaming piles of severe distortions and outright falsehoods, presented straight-faced alongside entirely accurate facts (Dickinson and Cowan 2023). They promote climate denial, creationism, and a “balanced” view of fascism itself. They lie outright about Black slavery, race science, and American history. Yet eight US states now officially approve “Prager University” videos in public schools and/or universities. Beyond that, they are widely viewed by huge numbers of right-wing and evangelical homeschoolers (Lewis 2018). At this writing, these videos have racked up tens of billions of views.
Seen as a principal source of “woke ideology,” the social sciences come in for particularly intense agnosis. The state of Florida has dictated that its twenty-eight public colleges may not “include a curriculum that teaches identity politics” or one that “is based on theories that systemic racism, sexism, oppression, and privilege are inherent in the institutions of the United States and were created to maintain social, political, and economic inequities” in their Introduction to Sociology courses, soon to be followed by similar ideological restrictions on American History (Whitford 2026). Other red states have adopted a myriad of strategies to prevent young people from being exposed to “woke” ideas.
Another factor eroding scientific norms is the elevation of populist pseudoscience to the highest levels of government. The current Secretary of Health and Human Services, Robert F. Kennedy Jr., is a notorious promoter of long-discredited fringe theories. His false belief that vaccines cause autism has underwritten many states’ relaxation of pediatric vaccination requirements and has already resulted in a large increase in the incidence of measles in the United States. In 2025, Kennedy dismissed all seventeen members of the CDC's vaccine advisory group, replacing them with vaccine deniers, including several with zero scientific qualifications of any kind. Kennedy also supported a revision of the food pyramid that encourages consumption of saturated fats such as beef tallow, long known to increase cardiovascular risk. At least eight US states have introduced or passed legislation to ban “chemtrails,” an old conspiracy theory about the contrails formed by water vapor in jet exhaust, which freezes as aircraft fly through cold air. Another concern is the proliferation of low-quality, poorly controlled studies of unregulated drugs and nutritional supplements for purposes ranging from body-building to cognitive enhancement, supported by the huge supplement industry and its lobbies.
Conclusions
What will be the consequences of American fascism for science, and especially for STS itself?
Reduced federal funding. Within the United States, federal funding for areas of science the Trump administration considers “woke” remains extremely precarious. Grant proposals for key focus areas of recent STS, such as gender, race, climate, environment, and decolonial studies, are unlikely to succeed. Trumpist political appointees have already denied numerous grants to STS researchers.
Full disclosure: my partner, Prof. Gabrielle Hecht, had a major NSF grant award first suspended, then withdrawn, in the first few days of the current administration. As a result, she has taken a post at Aix Marseille University in France as part of the “Safe Place for Science” (SPFP) program. We now live in Marseille as permanent residents of France; I retired from Stanford years before I originally intended in order to make the move possible. She is far from alone. Among the SPFP hires, I recently met astrophysicist Kartik Sheth, who lost his job as NASA's Associate Chief Scientist in March 2025 as a direct result of the Trump firings.
Reduced availability of federal data of all kinds, accompanied by reduced data quality and completeness.
Ideological distortion. Desperate for funding, scientists will be sorely tempted to twist their grant proposals, and even their research agendas, to “demonstrably advance the President's policy priorities” as now required by EO 14332. Capitulations will run the gamut from substituting euphemisms for prohibited terms to fundamental changes in orientation whereby researchers turn away from defunded areas such as climate change, vaccines, and cancer. Many of us will be tempted to do the same, simply to keep our careers and/or avoid the damaging consequences of outspoken resistance. I sympathize deeply with those of you in this position. As for me, I feel very lucky to have retired and left the United States, leaving me free to say exactly what I think.
Increased reticence of human subjects. Privacy and security concerns have grown among the many populations this administration seeks to suppress or harm. Federal intimidation will have its intended effect: reducing these groups’ willingness to speak openly. Researchers will need to be extremely careful to protect the identities of those who do agree, and find ways to document the views of those who seek to remain invisible.
Reduced academic freedom. We have already seen a dramatic reduction of academic freedom, especially with respect to issues of gender, race, immigration, and war crimes committed in Israel and Gaza. Many universities rapidly dismantled their existing diversity, equity, and inclusion (DEI) efforts under the administration's Sauron-like gaze. While I cannot say I will miss the tedious, bombastic DEI training videos my previous employer forced upon us, they probably served a useful purpose in raising awareness about on-the-ground realities of discrimination, prejudice, and power differentials.
Brain drain. In late 2025, about three-quarters of American scientists said they were looking to leave the country (Witze 2025), as I have. The coming brain drain from America to Europe, China, and elsewhere will resemble the exodus of European scientists to America in the 1930s—only larger. This path may especially appeal to younger scientists and scholars who find their career paths suddenly blocked, and I cannot but encourage it.
How can STS Meet this Moment?
As ethnographers, anthropologists, sociologists, and historians, we STS scholars are ourselves scientists. If you have not yet done so, I invite you to proudly claim that label. STS research has helped improve scientific practice. STS concepts such as coproduction, path dependence, and the social, ethical, and political consequences of classification have been taken up not only in the sciences, but throughout the academy.
So first and foremost, stay the course. Keep doing what we do best. As the scholar of fascism Timothy Snyder (2017) urges, do not obey in advance.
Next, we should make it our mission to track and analyze changes to the landscape of both natural and social sciences. How are knowledge infrastructures being reconfigured? What are the contests over legitimacy and standards of evidence? Are methods or data being revised, either in the open or in secret? What data are no longer being collected? Are historical data being preserved? If not, who is trying to recover them, and how can you help? The Environmental Data and Governance Initiative—begun in 2016 following Trump's first election and led, in part, by STS scholars—is one example of a valuable effort to maintain accountability and preserve data under ideological attack (Dillon et al. 2017, 2019; Alvarado Rojas et al. 2025).3
STS has a long, proud tradition of micro-studies and close ethnography. But in this changed environment, micro-studies alone are severely insufficient. Instead, we need what Gabrielle Hecht (2018) calls “interscalar methods.” To maintain ethnographic distance, we rightly shy away from normativity, and STS has always valued critical approaches and minority perspectives. But today, the old relativism and the murky language of high theory are dangerous luxuries. In the face of populist “common sense” science and the degradation of hard-won knowledge systems through paper mills, fraud, and untrustworthy AI, we must unequivocally defend fundamental norms of evidence, argument, credentialing, and method.
So much of what is happening now has happened before. We must re-engage the histories and geographies of science under fascism and autocracy. Lysenkoism and Nazi science are on everyone's lips, but other examples are much more recent and even contemporary: Franco's Spain. The Pinochet regime in Chile. Science in the PRC and post-Soviet Russia. Orban's Hungary. There are dissertations here.
In closing, I’ll admit that it is hard to find a bright spot in this dark landscape of intimidation and fear. But the antidote to agnosis, anxiety, and despair is action. Hope is work, and work brings hope. Beyond personal political action, more of us could take up the role of public intellectual. STS already has many brilliant examples: Naomi Oreskes, Alondra Nelson, Robert Proctor, Ben Franta, and Becca Lewis, to name just a few. Venues such as Aeon, The Conversation, The Guardian, and The Atlantic are well within reach. Professional societies of all sorts—4S among them—can play a role in facilitating far-reaching publications for broad audiences. Both faculty and students can re-learn writing, turning away from dry academic prose toward more engaging, funny, and trenchant styles.
Is nightfall coming? Is this the end of science? Maybe. But there is still time to defend it. With our allies and in our numbers, we can hold the line.
Author's Note
This text is a revised version of my Bernal Prize acceptance speech, delivered at the 4S meeting in Seattle in September 2025. I have checked my facts and included some citations to help readers find relevant information, but this is not intended as a scholarly article. No AI was used in researching or composing this piece.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.
Footnotes
1. Members of the Monitoring, Modeling, and Memory research group included previous Bernal Prize winners Geoffrey Bowker and Susan Leigh Star, as well as Christine Borgman, David Ribes, Steven Jackson, and numerous graduate students and postdocs.
2. Statement on home page, https://www.noaa.gov/climate.
3. EDGI is available at https://envirodatagov.org (accessed July 2026).
References
Alvarado Rojas Alejandro, Vera Lourdes, Hansen Steve, Nost Eric, Wylie Sara, Alder Cole. 2025. “Infrastructuring Care: Co-Designing Computational Notebooks for Environmental Data Justice.” Science, Technology, and Human Values. Online First: 1–32. https://doi.org/10.1177/01622439251356930
Google Scholar
Bernal J. D. 1939. The Social Function of Science, 415. London: George Routledge & Sons.
Google Scholar
Brembs Björn, Button Katherine, Munafò Marcus. 2013. “Deep Impact: Unintended Consequences of Journal Rank.” Frontiers in Human Neuroscience 7 (Article 291): 1–12. https://doi.org/10.3389/fnhum.2013.00291
PubMed
Google Scholar
Caudill David S. 2020. “Expertise in Political Contexts: Latour Avec the Third Wave in Science and Technology Studies.” Law, Technology and Humans 2 (2): 4–21. https://doi.org/10.5204/lthj.v2i2.1413.
Web of Science
Google Scholar
Chen Lulin, Li Yizhao, Wang Jie, Li Yue, Tan Xiaoli, Guo Xiaoyan. 2024. “Knowledge, Attitudes and Practices About Research Misconduct among Medical Residents in Southwest China: A Cross-Sectional Study.” BMC Medical Education 24 (1): 284. https://doi.org/10.1186/s12909-024-05277-6.
PubMed
Web of Science
Google Scholar
Collins Harry M., Evans Robert. 2017. Why Democracies Need Science. New York: Polity Press.
Google Scholar
Committee on Publication Ethics (COPE) International Association of Scientific, Technical & Medical Publishers (STM) 2022. Paper Mills—Research Report from COPE & STM. Committee on Publication Ethics. https://doi.org/10.24318/jtbG8IHL.
Google Scholar
Custer Sara. 2026. “How DOGE Gutted the NEH in 22 Days.” Inside Higher Ed. March 11. https://www.insidehighered.com/news/faculty-issues/research/2026/03/11/how-doge-gutted-neh-22-days.
Google Scholar
De Bell Garrett. 1970. The Environmental Handbook. New York: Ballantine Books.
Google Scholar
Dickinson Robert, Cowan Tom. 2023. “The Kids Are Alt-Right: An Introduction to PragerU and Its Role in Radicalization in the United States.” Patterns of Prejudice 56 (2–3): 95–122. https://doi.org/10.1080/0031322X.2023.2219167.
Google Scholar
Dillon Lindsey, Lave Rebecca, Mansfield Becky, et al. 2019. “Situating Data in a Trumpian Era: The Environmental Data and Governance Initiative.” Annals of the American Association of Geographers 109 (2): 545–55. https://doi.org/10.1080/24694452.2018.1511410.
Web of Science
Google Scholar
Dillon Lindsey, Walker Dawn, Shapiro Nicholas, et al. 2017. “Environmental Data Justice and the Trump Administration: Reflections from the Environmental Data and Governance Initiative.” Environmental Justice 10 (6): 186–92. https://doi.org/10.1089/env.2017.0020.
Web of Science
Google Scholar
Edwards Paul, Jackson Steven J., Chalmers Melissa K., et al. 2013. Knowledge Infrastructures: Intellectual Frameworks and Research Challenges. Ann Arbor: Deep Blue. http://hdl.handle.net/2027.42/97552.
Google Scholar
Edwards Paul N. 2010. A Vast Machine: Computer Models, Climate Data, and the Politics of Global Warming. MIT Press.
Google Scholar
Edwards Paul N. 2012. “Entangled Histories: Climate Science and Nuclear Weapons Research.” Bulletin of the Atomic Scientists 68 (4): 28–40. https://doi.org/10.1177/0096340212451574.
Web of Science
Google Scholar
Edwards Paul N. 2016. “Trump Questionnaire Recalls Dark History of Ideology-Driven Science.” The Conversation, https://doi.org/10.64628/AAI.436u3qvc9. December 14.
Google Scholar
Edwards Paul N. 2019. “Knowledge Infrastructures Under Siege: Climate Data as Memory, Truce, and Target.” In Data Politics: Worlds, Subjects, Rights, edited by Bigo Didier, Isin Engin, Ruppert Evelyn. Routledge Studies in International Political Sociology. 21–42. New York: Routledge.
Crossref
Google Scholar
Edwards Paul N., Bowker Geoffrey C., Jackson Steven J., Williams Robin. 2009. “Introduction: An Agenda for Infrastructure Studies.” Journal of the Association for Information Systems 10 (5): 6. https://doi.org/10.17705/1jais.00200
Web of Science
Google Scholar
Edwards Paul N., Jackson Steven J., Bowker Geoffrey C., Knobel Cory P. 2007. Understanding Infrastructure: Dynamics, Tensions, and Design. Final report of the workshop, "History and Theory of Infrastructure: Lessons for New Scientific Cyberinfrastructures." Deep Blue. http://hdl.handle.net/2027.42/49353.
Google Scholar
Franta Benjamin. 2021a. “Early Oil Industry Disinformation on Global Warming.” Environmental Politics 30 (4): 663–68. https://doi.org/10.1080/09644016.2020.1863703.
Web of Science
Google Scholar
Franta Benjamin. 2021b. “Weaponizing Economics: Big Oil, Economic Consultants, and Climate Policy Delay.” Environmental Politics 30 (4): 1–21. https://doi.org/10.1080/09644016.2021.1947636
Google Scholar
Hecht Gabrielle. 2018. “Interscalar Vehicles for an African Anthropocene: On Waste, Temporality, and Violence.” Cultural Anthropology 33 (1): 1. https://doi.org/10.14506/ca33.1.05.
Web of Science
Google Scholar
Latour Bruno. 2018. Down to Earth: Politics in the New Climatic Regime. New York: Polity. Translated by Catherine Porter.
Google Scholar
Lewis Rebecca. 2018. Alternative Influence: Broadcasting the Reactionary Right on YouTube. White Paper. New York: Data & Society. https://datasociety.net/wp-content/uploads/2026/02/DS_Alternative_Influence.pdf
Crossref
Google Scholar
Liang Weixin, Zhang Yaohui, Wu Zhengxuan, et al. 2025. “Quantifying Large Language Model Usage in Scientific Papers.” Nature Human Behaviour 9: 2599–2609. https://doi.org/10.1038/s41562-025-02273-8
PubMed
Web of Science
Google Scholar
MacGillis Alec. 2025. “Trump’s War on Measurement Means Losing Data on Drug Use, Maternal Mortality, Climate Change and More.” ProPublica April 18. https://www.propublica.org/article/trump-doge-data-collection-hhs-epa-cdc-maternal-mortality.
Google Scholar
Martin Brian. 1989. “The Sociology of the Fluoridation Controversy: A Reexamination.” The Sociological Quarterly 30 (1): 59–76. https://doi.org/10.1111/j.1533-8525.1989.tb01511.x
Web of Science
Google Scholar
Martin Brian. 1991. Scientific Knowledge in Controversy: The Social Dynamics of the Fluoridation Debate. Albany: State University of New York Press.
Google Scholar
Mervis Jeffrey, Hersher Monica. 2026. “The Great Government Brain Drain.” Science 391 (6784): 428–29. https://doi.org/10.1126/science.aef8893.
PubMed
Web of Science
Google Scholar
Naddaf Miryam, Quill Elizabeth. 2026. “Hallucinated Citations Are Polluting the Scientific Literature. What Can Be Done?” Nature 652 (8108): 26–9. https://doi.org/10.1038/d41586-026-00969-z.
PubMed
Google Scholar
Oreskes Naomi, Conway Erik M. 2010. Merchants of Doubt: How a Handful of Scientists Obscured the Truth on Issues from Tobacco Smoke to Global Warming. New York: Bloomsbury Press.
Google Scholar
Plumer Brad, Dzombak Rebecca. 2025. “All Authors Working on Flagship US Climate Report Are Dismissed. Climate.” The New York Times April 28. https://www.nytimes.com/2025/04/28/climate/national-climate-assessment-authors-dismissed.html.
Google Scholar
Proctor Robert N. 2012. Golden Holocaust: Origins of the Cigarette Catastrophe and the Case for Abolition. Berkeley: University of California Press. https://doi.org/10.1525/9780520950436.
Google Scholar
Proctor Robert N., Schiebinger Londa. 2008. Agnotology: The Making and Unmaking of Ignorance. Stanford: Stanford University Press.
Google Scholar
Richardson Reese. 2025. A Do-or-Die Moment for the Scientific Enterprise. Blog entry, August 4. https://reeserichardson.blog/2025/08/04/a-do-or-die-moment-for-the-scientific-enterprise/.
Crossref
Google Scholar
Richardson Reese A. K., Hong Spencer S., Byrne Jennifer A., Stoeger Thomas, Amaral Luís A. Nunes. 2025. “The Entities Enabling Scientific Fraud at Scale Are Large, Resilient, and Growing Rapidly.” Proceedings of the National Academy of Sciences 122 (32): e2420092122. https://doi.org/10.1073/pnas.2420092122.
Web of Science
Google Scholar
Schneider Stephen H. 2010. Science as a Contact Sport: Inside the Battle to Save Earth’s Climate. Washington, DC: National Geographic.
Google Scholar
Snyder Timothy. 2017. On Tyranny: Twenty Lessons from the Twentieth Century. New York: Crown.
Google Scholar
Whitford Emma. 2026. “Florida Hands Down Sociology Curriculum to State Colleges.” Inside Higher Ed. February 20. https://www.insidehighered.com/news/faculty-issues/curriculum/2026/02/20/florida-hands-down-sociology-curriculum-state-colleges.
Google Scholar
Witze Alexandra. 2025. “75% of US Scientists Who Answered Nature Poll Consider Leaving.” Nature 640 (8058): 298–99. https://doi.org/10.1038/d41586-025-00938-y.
PubMed
Web of Science
Google Scholar
Yourish Karen, Daniel Annie, Datar Saurabh, White Isaac, Gamio Lazaro. 2025. “These Words Are Disappearing in the New Trump Administration.” The New York Times March 7. https://www.nytimes.com/interactive/2025/03/07/us/trump-federal-agencies-websites-words-dei.html.
Google Scholar
Biographies
Paul N. Edwards is a Professor (Emeritus) of Information, History, and STS. His areas of research are the history, politics, and culture of information technology and climate science.
No comments:
Post a Comment