In January 2025, a Chinese company called DeepSeek released an AI model that matched the performance of leading American systems at a fraction of the cost. Analysts immediately reached for a familiar metaphor: this was America’s new “Sputnik moment” (Institute of Geoeconomics, 2025). The comparison holds, though the satellite was always the least important part of that story. What mattered was what the United States chose to do with its people once it felt threatened.
In 1958, the United States answered a technological shock with the largest federal investment in education in its history, and that investment was deliberately broad. It funded physicists, yes, but also linguists, historians, area-studies scholars, school counselors, and eventually the humanities as a whole. The bet was that an educated citizenry, across every field of knowledge, would prove to be the country’s deepest source of strength. That bet paid dividends for half a century.
Today, facing a genuinely comparable challenge, we are making the opposite bet. We are dismantling the federal education infrastructure, canceling research, driving scholars overseas, and treating broad education as a luxury or even a threat. If the Cold War teaches us anything, it is that this is precisely backwards.
The shock before Sputnik
Eight years earlier, on August 29, 1949, the Soviet Union detonated its first atomic bomb at the Semipalatinsk test site, years ahead of American intelligence estimates, which had generally placed a Soviet bomb in the mid-1950s (National Security Archive, 2019). Losing the nuclear monopoly was bad enough. Losing it ahead of schedule meant the Soviet scientific establishment was far more capable than Americans had allowed themselves to believe. The Cold War stopped being a contest the United States had already won technologically and became an open-ended race between two systems for producing trained minds.
The institutional response began almost immediately. The loss of the monopoly was a direct stimulus for NSC-68 in April 1950, the strategy document that framed the conflict as a mobilization of total national capabilities rather than a comparison of arsenals (National Security Archive, n.d.). The National Science Foundation, which had been stalled in Congress for five years after Vannevar Bush proposed it in 1945, was finally signed into law in May 1950, eight months after President Truman announced the Soviet test (National Science Foundation, n.d.). That signature, eight years before the NDEA, marks the true beginning of federal responsibility for the nation’s scientific talent pipeline. Through the following decade, a steady stream of studies asked whether the Soviets could out-produce America in scientists and engineers, and most answered yes. The National Research Council published Soviet Professional Manpower in 1955, Nicholas DeWitt’s Harvard analyses of Soviet education circulated through Washington, and President Eisenhower convened his Committee on Education Beyond the High School in 1956, a full year before Sputnik flew (National Academies Press, 1955).
The 1949 test also previewed a darker possibility, because the same shock that built institutions nearly tore them down. The revelation that espionage had accelerated the Soviet program fed McCarthyism, loyalty oaths, and purges that drove scholars out of American universities in the early 1950s. For several years the United States ran both responses at once, investing in education with one hand while attacking educators with the other. The constructive response won out, but never cleanly, and the two impulses were never fully separate. Even the NDEA carried the suspicion of the era inside it: every student who took one of its loans had to sign an affidavit disclaiming belief in or support of any subversive organization, a requirement so corrosive to academic freedom that Harvard, Yale, Princeton, and more than a hundred other institutions refused to participate until Congress repealed the provision in 1962 (Harvard Crimson, 2012). An external technological shock produces fear long before it produces policy, and a society chooses what to build with that fear.
The first Sputnik moment
When the Soviet Union launched Sputnik 1 on October 4, 1957, the American public reaction bordered on panic. Beneath the military fear ran a deeper one, that Soviet schools were producing better minds and that the United States was falling behind in the long competition between two ways of organizing a society. But if Sputnik was the spark, the fuel had been accumulating since 1949. The manpower studies, the federal machinery, and even the policy arguments were already in place, which is why Congress could respond within a year by passing the National Defense Education Act of 1958, the first comprehensive federal commitment to American education (U.S. Senate, n.d.).
The NDEA is usually remembered as a science bill, and it did pour money into mathematics and science teaching. But its architects understood national defense in strikingly human terms. The act created low-interest student loans so that talented students of modest means could attend college at all. It funded graduate fellowships to rebuild the professoriate. And under Title VI, it built nineteen language and area-studies centers at American universities, on the theory that the United States could not lead a divided world if its citizens could not understand the languages, histories, and cultures of other peoples (Legters, 1965). Anthropologists, political scientists, and historians of Asia, Africa, and Eastern Europe were funded as instruments of national security, because understanding people was understood to be a security capability.
Anthropology should be honest about the price of that bargain. The same security money that built the field also entangled it with the agencies that paid, and the entanglement ended badly. Project Camelot, a 1964 Army project that recruited social scientists to model revolution and counterinsurgency in Latin America, collapsed in international scandal and pushed the discipline into an ethical reckoning that has never fully ended (Price, 2016). The funding bought breadth, and it also bought influence over what got studied and for whom. That tension sits inside any argument, including this one, that asks the security state to pay for learning.
The results compounded for decades. Enrollment was already climbing on the strength of the GI Bill, the baby boom, and postwar prosperity, and the new federal commitment accelerated the climb: college enrollment grew from 3.6 million students in 1959 to 8.6 million in 1970 (NCES, 2013). The NDEA’s framework flowed directly into the Higher Education Act of 1965, which created the need-based aid system that still exists today (Congressional Research Service, 2020), and into the founding of the National Endowment for the Humanities that same year (NEH, n.d.). No single statute produced what came next, but the generation educated inside this system staffed the laboratories, firms, and universities that gave the country the semiconductor industry, the internet, the modern biomedical enterprise, and a research establishment the rest of the world spent fifty years trying to copy. A national security scare became, almost by accident, one of the great engines of social mobility in American history.
Clarence Randall and the myth of the specialist
It would be easy to assume that the business establishment of the 1950s wanted narrow technical training and nothing more. The career of Clarence B. Randall demonstrates otherwise. Randall was about as establishment as an American could be: president and then chairman of Inland Steel, special assistant to President Eisenhower, and chairman of the Commission on Foreign Economic Policy (Eisenhower Presidential Library, n.d.; Wikipedia, n.d.). He spent the height of the Cold War advising the government on how American industry could outcompete the Soviet bloc.
And yet the argument Randall kept returning to in public, in books and in lectures with titles like A Businessman Looks at the Liberal Arts, was the case for breadth. In his 1961 book The Folklore of Management, written at the very moment the nation was rushing to produce technical specialists, Randall attacked the over-specialized executive as a myth of modern business. He warned that “galloping specialization can bring any company to the brink of chaos” and insisted that organizations needed leaders with “the breadth of vision only a liberal education can provide” (Randall, 1961). A steel executive and Cold War policymaker, looking out at the same technological race that produced the NDEA, concluded that the specialist who knew only his specialty was a liability, and that the future belonged to people educated broadly enough to see whole systems, weigh competing values, and adapt when the world changed.
Randall did not speak for everyone. Admiral Hyman Rickover, the father of the nuclear navy, spent the same years demanding a leaner and relentlessly technical school system, and his Education and Freedom appeared in 1959 to wide acclaim (Rickover, 1959). The Cold War education debate was a genuine fight between breadth and rigor-as-specialization. What matters is how the fight resolved: the legislation that emerged split the difference in breadth’s favor, funding physics fellowships alongside Russian language programs, student loans for every field of study, and, within a few years, the humanities endowment. The working consensus held that scientific power and humanistic understanding reinforced each other, and that a free society defends itself by educating whole human beings.
What anthropology knows about education
My own discipline has something distinctive to say here, because anthropologists study education as something much larger than the transfer of job skills. Margaret Mead, writing in the middle of the Second World War, observed that every human society educates its young, but that societies differ profoundly in what they believe education is for. Some treat it as the transmission of a fixed tradition. Others treat it as the deliberate creation of citizens capable of living in a world their parents could not have imagined (Mead, 1943). Mead was no cheerleader for American schooling, which she thought too often expressed the will to teach, convert, and colonize rather than the will to learn, but the second vision is the aspiration worth recovering from her essay. Education, in the anthropological sense, is enculturation. It is the process by which a society reproduces itself, and what a society chooses to teach is a statement about who it intends to become.
Pierre Bourdieu offered a harder lesson. Schools, he argued, reward the cultural capital children inherit from their families, the knowledge, dispositions, and ways of speaking and reasoning that the educated pass down, and then disguise that inheritance as individual merit. Left to itself, the education system reproduces inequality across generations while appearing scrupulously fair (Bourdieu & Passeron, 1977). Bourdieu would have scoffed at any claim that the NDEA cured this, and he would have been right to, because the reproductive machinery he described still runs in every admissions office. What the postwar programs did was force the gates open wider. Loans and need-based grants put a university education, and the cultural capital that travels with it, within reach of steelworkers’ children who would never otherwise have entered, and the income data show what followed. Workers with a bachelor’s degree today earn roughly two thirds more per week than workers with only a high school diploma, and they experience markedly lower unemployment at every point in the economic cycle (U.S. Bureau of Labor Statistics, 2025). Credential inflation is real, and a degree guarantees nobody anything, but few public investments have a comparable record of moving families upward.
Anthropology also insists on holism, the methodological commitment to studying human life as an integrated whole rather than as isolated variables. Jean Lave and Etienne Wenger showed that real learning is situated, embedded in communities of practice, in relationships between people who know and people who are coming to know (Lave & Wenger, 1991). Learning is something people do together. This matters enormously for how we think about AI, because a language model does enculturate its users into something, and we should be precise about what. Hours spent with a fluent answer machine apprentice a young person into the habits of querying and accepting. They do little to apprentice her into a community of scientists, teach the embodied judgment of a field archaeologist, or model what intellectual honesty looks like when the evidence turns against a hypothesis. Those things are transmitted between humans, in institutions built for the purpose. When we defund the institutions, we sever the transmission.
The dividends of breadth
Pull these threads together and the case for broad education rests on claims the Cold War record supports.
Broad education builds critical thinking. Specialists trained only in their specialty are dependent on others to frame their problems. A person educated across history, science, literature, and quantitative reasoning has multiple frameworks to test against each other, which is the substance of critical thought. This was Randall’s point about executives, and it applies with greater force to citizens evaluating political claims, scientific findings, or the output of an AI model.
It also exposes people to what they would never otherwise encounter. The farm kid who takes a required anthropology course and the engineering student who stumbles into a Russian literature seminar are doing the quiet work through which a large, plural society generates mutual intelligibility among its members. Narrow vocational training eliminates exactly these encounters.
The economic case is just as strong. Education raises earnings, lowers unemployment, and, in Bourdieu’s terms, pries open access to the cultural capital that gates economic opportunity. The postwar expansion of higher education helped build the broadest middle class the country has ever had.
Broad education strengthens society, and a strong society is where innovation comes from. Democracies run on a shared capacity to deliberate, to distinguish evidence from assertion, and to imagine the lives of fellow citizens unlike ourselves, and Mead understood schooling as the deliberate formation of such citizens. A society that stops forming them does not remain a deliberative society for long. The same investment drives discovery, because the breakthroughs of the past seventy years came overwhelmingly from people trained in the federally supported university system, and many came from unplanned collisions between fields. The biotechnology revolution grew out of basic science nobody could justify on quarterly returns, and the early internet out of a research culture paid to follow open-ended questions. Breadth is where the unexpected combinations come from.
And broad education strengthens national security, in exactly the way the NDEA’s drafters understood. Security rests on the linguistic and cultural knowledge to understand rivals and allies, the scientific depth to maintain technological advantage, and the civic resilience to resist manipulation, as much as it rests on weapons. Title VI funded area studies because the government knew it could not afford strategic ignorance of other societies. That need persists today, and it now extends to understanding the society on the other side of the AI race.
The reversal
Now consider what the United States is actually doing at its second Sputnik moment.
In March 2025, the administration issued an executive order directing the dismantling of the Department of Education, and within a year the department’s workforce had been cut nearly in half (NEA, 2026). More than a hundred education research contracts, collectively worth around a billion dollars, were terminated, and almost no new federal education research was commissioned in 2025, which means a generation of evidence about what actually works in schools may simply never exist (Hechinger Report, 2026). In May 2026, an additional $2 billion in education grants was withheld (Education Week, 2026).
The research enterprise is contracting with it. The National Science Foundation and the National Institutes of Health each awarded roughly a quarter fewer grants in 2025 than their decade average (U.S. News, 2026). Brown alone paused doctoral admissions in at least six departments for the coming academic year, among them anthropology, classics, and several language programs, the very fields the Cold War generation deliberately built (Brown Daily Herald, 2025). And the human consequence is already measurable. In a 2025 Nature poll of more than 1,600 scientists, a self-selected sample but a striking one, three quarters said they were considering leaving the country, and applications from U.S.-based researchers to European Research Council early-career grants have nearly tripled (STAT, 2025) as Europe runs explicit recruitment campaigns like the half-billion-euro Choose Europe for Science initiative (European Commission, 2025; AAU, 2025). During the Cold War, the United States was the destination for the world’s displaced scientific talent, and it harvested the dividends for generations. We are now exporting that talent voluntarily.
The standard defense of these policies is that they target a bureaucracy rather than education itself, and that schooling rightly belongs to states and families. The argument deserves a straight answer, and the answer is that the cuts are landing somewhere else. Canceled research grants, withheld student aid, shuttered doctoral programs, and an exodus of working scientists are losses to education under any theory of who should administer it. A federalist who wants states to run their schools still needs the research, the aid dollars, and the scientists to exist.
Measure this against the dividends of breadth and the damage is systematic. Cutting education research degrades our ability to teach critical thinking, because we stop learning how to teach at all. Eliminating humanities and social science programs eliminates exposure. Withholding grant aid and raising the cost of college restores the pre-1958 world in which cultural capital stays where it already is, and poverty hardens. A society that defunds its own deliberative capacity weakens itself. An innovation system bleeding researchers to Europe and China innovates less. And a nation that abandons area studies, language training, and the scientific workforce in the middle of a strategic competition has confused austerity with strength.
The official response to the AI challenge makes the contrast with 1958 painfully clear. The administration’s signature education initiative, an April 2025 executive order on advancing AI education for American youth, directs agencies to promote AI literacy in K-12 schools through task forces and public-private partnerships (The White House, 2025). Set aside its merits and notice its shape. Where the NDEA answered Sputnik with student loans, graduate fellowships, teacher training, language centers, and counseling across the whole of education, the current response is a single narrow program teaching children to use the technology of the moment, pursued simultaneously with the demolition of everything around it. It is as if the country had answered Sputnik by teaching schoolchildren to operate telescopes while closing the universities. China, meanwhile, is running the narrow version of the race with total commitment. Chinese state media report that Beijing now requires AI instruction across its schools as part of a nationwide rollout (Xinhua, 2025), and the domestic university pipeline already produces far more STEM doctorates than ours does (Center for Security and Emerging Technology, 2021). The United States is thus conceding ground in the narrow contest while dismantling the broad capacity that was always its real advantage.
This is the 1949 pattern returning. A technological shock from a rival power has once again produced two simultaneous responses, one that would invest in the nation’s minds and one that attacks the institutions that train them. In the early 1950s the constructive response ultimately prevailed over the purges, and the country collected the dividends for fifty years. This time the destructive response is the one with momentum, and the dividends will accrue to whoever offers our researchers a home.
Educating humans in the age of thinking machines
There is a deeper error underneath the policy choices, and it is anthropological. The implicit theory of the moment is that intelligence is becoming a commodity, that machines will soon supply the thinking, and that human education can therefore narrow to whatever the machines cannot yet do. This gets the relationship exactly wrong.
AI raises, rather than lowers, the educational demands on a citizenry. A society saturated with fluent, persuasive, machine-generated text needs citizens who can evaluate claims independently, who know enough history to recognize a manipulated narrative, enough statistics to interrogate a confident assertion, and enough about how knowledge is made to ask what a model’s answer is actually based on. Critical thinking about AI draws on epistemology, history, psychology, and ethics far more than it draws on any technical skill, which is to say it draws on broad education. A student who has only been taught to prompt a model has been enculturated into dependence on it, and that is a very different thing from being educated about it.
Anthropology also reminds us that every technology is embedded in a culture, and that the AI race is, beneath the benchmarks, a competition between societies with different ideas about what people are for. That was true of the Cold War too, and it is why the American response in 1958 was, at bottom, a humanist one. Rather than imitate the Soviet model and conscript narrow technical cadres, the United States bet on breadth, on access, and on the unpredictable productivity of millions of broadly educated free people. The bet paid off in science, in prosperity, and in the legitimacy of the society itself.
There is a trap inside this essay’s own argument, and naming it is better than hiding it. Justifying education by the AI race is an instrumental argument, and instrumental arguments are how education gets narrowed. The NDEA’s drafters wrote their purpose clause in exactly those terms, promising trained manpower sufficient “to meet the national defense needs of the United States,” and the breadth the act funded was partly a happy accident of coalition politics (Wikipedia, n.d.). Security framing is what moves money through Congress. It has never been the reason education matters, and a case built only on beating China will last exactly as long as the fear of China does. If the AI race frightens the country back into investing in its schools, universities, and libraries, take the win, because the investment will pay off in all the ways this essay has catalogued. The lasting reasons run deeper, and they are the ones Mead and Randall understood: a society that educates its people broadly is choosing what kind of people it intends to have.
Clarence Randall’s phrase deserves the last word, because it names exactly what is at stake. The breadth of vision only a liberal education can provide was, in his telling, a business necessity. In ours, it is a civilizational one. The models will keep improving regardless of what we do. The open question, the one the Cold War generation answered so much better than we currently are, is whether the humans will.
References
Association of American Universities. (2025). Scientific talent in America: Going abroad or choosing not to come. https://www.aau.edu/key-issues/scientific-talent-america-going-abroad-or-choosing-not-come
Bourdieu, P., & Passeron, J.-C. (1977). Reproduction in education, society and culture. Sage. https://www.sagepub.com/en-us/nam/reproduction-in-education-society-and-culture/book203162
Brown Daily Herald. (2025, October). PhD admissions paused in at least six humanities, social science departments. https://www.browndailyherald.com/article/2025/10/phd-admissions-paused-in-at-least-six-humanities-social-science-departments
Center for Security and Emerging Technology. (2021). China is fast outpacing U.S. STEM PhD growth. Georgetown University. https://cset.georgetown.edu/publication/china-is-fast-outpacing-u-s-stem-phd-growth/
Congressional Research Service. (2020). International education programs in the Higher Education Act (R46508). https://www.congress.gov/crs-product/R46508
Dwight D. Eisenhower Presidential Library. (n.d.). Randall, Clarence B.: Papers. https://www.eisenhowerlibrary.gov/research/finding-aids/randall-clarence-b
Education Week. (2026, May). Trump holds back $2 billion for education grants. What will happen next? https://www.edweek.org/policy-politics/trump-holds-back-2-billion-for-education-grants-what-will-happen-next/2026/05
European Commission. (2025, May 23). Choose Europe for Science: EU comes together to attract top research talent. https://research-and-innovation.ec.europa.eu/news/all-research-and-innovation-news/choose-europe-science-eu-comes-together-attract-top-research-talent-2025-05-23_en
Harvard Crimson. (2012, May 21). NDEA grants ignite debate over Cold War loyalty. https://www.thecrimson.com/article/2012/5/21/ndea-grants-loyalty-oath/
Hechinger Report. (2026). How Trump 2.0 upended education research and statistics in one year. https://hechingerreport.org/proof-points-trump-upended-education-research-2025/
Institute of Geoeconomics. (2025). A “Sputnik” moment in the global AI race. https://instituteofgeoeconomics.org/en/research/2025070301/
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral participation. Cambridge University Press. https://www.cambridge.org/core/books/situated-learning/6915ABD21C8E4619F750A4D4ACA616CD
Legters, L. H. (1965). NDEA language and area centers: A report on the first five years. U.S. Office of Education. https://files.eric.ed.gov/fulltext/ED051683.pdf
Mead, M. (1943). Our educational emphases in primitive perspective. American Journal of Sociology, 48(6), 633-639. https://doi.org/10.1086/219260
National Academies Press. (1955). Soviet professional manpower: Its education, training, and supply. National Research Council. https://nap.nationalacademies.org/catalog/20224/soviet-professional-manpower-its-education-training-and-supply
National Center for Education Statistics. (2013). Total fall enrollment in degree-granting postsecondary institutions: Selected years, 1947 through 2023 (Table 303.10). Digest of Education Statistics. https://nces.ed.gov/programs/digest/d13/tables/dt13_303.10.asp
National Education Association. (2026). The plan to abolish the Education Department, one year later. https://www.nea.org/nea-today/all-news-articles/plan-abolish-education-department-one-year-later
National Endowment for the Humanities. (n.d.). About NEH. https://www.neh.gov/about
National Science Foundation. (n.d.). A timeline of NSF history. https://www.nsf.gov/about/history
National Security Archive. (2019, September 9). Detection of the first Soviet nuclear test, September 1949. https://nsarchive.gwu.edu/briefing-book/nuclear-vault/2019-09-09/detection-first-soviet-nuclear-test-september-1949
National Security Archive. (n.d.). U.S. intelligence and the detection of the first Soviet nuclear test, September 1949. https://nsarchive2.gwu.edu/nukevault/ebb286/index.htm
Price, D. H. (2016). Cold War anthropology: The CIA, the Pentagon, and the growth of dual use anthropology. Duke University Press. https://www.dukeupress.edu/cold-war-anthropology
Randall, C. B. (1961). The folklore of management. Little, Brown. https://books.google.com/books/about/The_Folklore_of_Management.html?id=JgxEAAAAIAAJ
Rickover, H. G. (1959). Education and freedom. E. P. Dutton. https://archive.org/details/educationfreedom00rick
STAT. (2025, December 17). Brain drain: Many scientists see better research options overseas. https://www.statnews.com/2025/12/17/research-cuts-fuel-scientific-brain-drain-american-science-shattered/
U.S. Bureau of Labor Statistics. (2025). Education pays, 2024. Career Outlook. https://www.bls.gov/careeroutlook/2025/data-on-display/education-pays.htm
U.S. News & World Report. (2026, January 2). The biggest developments in higher education policy in 2025. https://www.usnews.com/news/education-news/articles/2026-01-02/the-biggest-developments-in-higher-education-policy-in-2025
U.S. Senate. (n.d.). Sputnik spurs passage of the National Defense Education Act. https://www.senate.gov/artandhistory/history/minute/Sputnik_Spurs_Passage_of_National_Defense_Education_Act.htm
The White House. (2025, April 23). Advancing artificial intelligence education for American youth. https://www.whitehouse.gov/presidential-actions/2025/04/advancing-artificial-intelligence-education-for-american-youth/
Wikipedia. (n.d.). Clarence B. Randall. https://en.wikipedia.org/wiki/Clarence_B._Randall
Wikipedia. (n.d.). National Defense Education Act. https://en.wikipedia.org/wiki/National_Defense_Education_Act
Xinhua. (2025, December 24). China focus: AI takes root in basic education, nurturing future innovators. https://english.news.cn/20251224/55f17a6865cb40c58e0c48c81e8cd597/c.html