On December 26, 2004, the geological plates beneath Sumatra unleashed the third-most-powerful earthquake ever recorded. A gargantuan column of water raced toward Sri Lanka, India, Thailand, and Indonesia. None of these countries had advance-warning systems in place, so no one had time to prepare before the surge hit. Some 228,000 people died—the highest toll of any natural disaster so far this century.

Setting up prevention systems would have been inexpensive, especially compared with the countless billions the tsunami ultimately cost. But governments typically spend money on preventing disasters only after disasters strike, and the affected countries hadn’t experienced a major tsunami in years. After the events of 2004, USAID spent a tiny fraction of its budget to help fund an advance-detection system for the Pacific, which might have saved hundreds of thousands of lives had it been in place sooner. But some people would have seen such an investment as a “waste”—inefficient spending that could have gone toward some more immediate or tangible end.

DOGE has turned this dangerously flawed view into a philosophy of government. Last week, Elon Musk’s makeshift agency fired one of the main scientists responsible for providing advance warning when the next tsunami hits Alaska, Hawaii, or the Pacific Coast. The USAID document that describes America’s efforts to protect coastlines from tsunamis, titled “Pounds of Prevention”—riffing on the adage that an ounce of prevention is worth a pound of cure—now redirects to an error message: “The resource you are trying to access is temporarily unavailable.”

More than 800 workers at the National Oceanic and Atmospheric Administration have lost their job in recent weeks, including many who helped mitigate climate disasters, track hurricanes, predict ever-stronger storms, and notify potential victims. Meanwhile, cuts to volcano monitoring are crippling the government’s ability to measure eruption risk. DOGE is also reportedly preparing to cancel the lease on the government’s “nerve center” for national weather forecasts.

Musk has categorized as superfluous a good deal of spending that actually makes the country more resilient, at a time when catastrophic risk is on the rise. We never see the crises that the government averts, only the ones it fails to prevent. Preparing for them may seem wasteful—until suddenly, tragically, it doesn’t.

[Read: The diseases are coming]

The modern, globalized world is the most complex and interconnected environment that humans have ever navigated. That’s why the potential for catastrophic risk—that is, the risk of low-probability but highly destructive events—has never been greater. A single person getting sick can derail the lives of billions. A crisis in one country’s banking sector can crash economies thousands of miles away. Now is precisely the time when governments must invest more heavily in making themselves resilient to these kinds of events. But the United States is doing the opposite.

Donald Trump made the same mistake in his first term. In September 2019, his administration quietly eliminated an initiative that it saw as government waste: a $200 million program that tracked novel coronaviruses around the world. Three months later, COVID-19 infected its first victim in Wuhan. The U.S. government spent an estimated $4.6 trillion in response to the pandemic that emerged from that virus—roughly 23,000 times the budget for the preparedness program that could have helped mitigate its effects.

Complex systems—say, health care, or government, or industrial supply chains—without any built-in slack or redundancy are efficient but fragile. The effects of any disruption quickly cascade, and the potential for catastrophic risk grows. In 2021, a gust of wind turned a boat sideways in the Suez Canal—and upended the global economy, inflicting tens of billions of dollars in economic damage. Last year’s CrowdStrike outage is another example of an avalanche created by a minor problem within a system that was not resilient.

DOGE is courting these kinds of risks by automatically assuming that programs with no immediately obvious function—or at least none that Musk and his minions can discern—are wasteful. Some of its cost cutting may be eliminating genuine waste; no government spends its money perfectly. But DOGE’s campaign is riddled with errors, at the level of both understanding and execution. The agency’s strategy is akin to a climber replacing sturdy rope with low-cost string: We may not realize the full danger until it snaps.

Musk developed DOGE’s playbook when he took over Twitter, where resilience matters much less than it does in government. Gutting the social-media platform may have resulted in more harmful content and some outages, including one this week, but the stakes were low compared with the crucial government services that Musk is currently cutting. When X fails, memes go unposted. When the government fails, people can die.

The risks are not only to Americans but also to humanity, as technology and climate change have linked the destinies of far-flung people more closely and increased the likelihood of extinction-level calamities. It is not reassuring in this regard that Trump controls the world’s largest nuclear arsenal and that DOGE accidentally fired key people who manage it, that Trump doesn’t believe in climate change and is having Musk slash seemingly every agency designed to mitigate it, and that Musk summarized his view of AI risk by telling Joe Rogan that it presents “only a 20 percent chance of annihilation.” The United Nations Office for Disaster Risk Reduction—an organization that DOGE would certainly eliminate if it could—came up with a more sophisticated figure in 2023: By its estimate, there is a 2 to 14 percent chance of an extinction-level event in the 21st century. This is not a world in which the government should be running itself on a just-in-time basis.

Musk may flippantly acknowledge the risk in interviews, but DOGE’s fundamental ethos—Silicon Valley will fix what the government cannot—almost entirely ignores it.

[Read: The dictatorship of the engineer]

Americans can’t rely on Meta, Google, and Apple to build tsunami-early-warning systems, mitigate climate change, or responsibly regulate artificial intelligence. Preventing catastrophic risk doesn’t increase shareholder value. The market will not save us.

As DOGE hollows out the Federal Aviation Administration, fires extreme-weather forecasters, and implodes the National Institutes of Health, Americans are left to wonder: What happens when another plane crashes, or a hurricane hits Florida without sufficient warning, or the next pandemic takes America by surprise? Many people may die avoidable deaths for the rest of us to learn that one billionaire’s “waste” is really a country’s strength.

The first time Donald Trump threatened to use the power of the presidency to punish a university, I was the target. At UC Berkeley, where I was chancellor, campus police had at the last moment canceled an appearance by Milo Yiannopoulos, the alt-right political pundit who was then a star at Breitbart News, because of a violent attack on the venue by a group of outside left-wing activists who objected to Yiannopoulos’s presence. In the end, although these protesters caused significant damage both on campus and to shops and businesses in downtown Berkeley, the police restored peace. Yiannopoulos was safely escorted back to his hotel, where he promptly criticized the university for canceling his speech. But on the morning of February 2, 2017, I awoke to a tweet reading: “If U.C. Berkeley does not allow free speech and practices violence on innocent people with a different point of view - NO FEDERAL FUNDS?”

I didn’t worry much about Trump’s threat at the time. I now realize that was a mistake. American universities did not cause the onslaught that the second Trump administration is unleashing upon them. But they would be in a much stronger position today if they had made a proactive case to the public for their own importance—and taken steps to address their very real shortcomings.

In the aftermath of the Yiannopoulos episode and Trump’s tweet, I worried less about the potential loss of federal funding than about the enormous costs of hiring additional police and converting the campus into a riot zone over and over. Berkeley’s commitment to free speech all but guaranteed that more conflict was in store. Yiannopoulos had announced that he would come back, and Ann Coulter soon accepted an invitation to speak at Berkeley as well. For a time, my concerns seemed justified. Berkeley spent millions of dollars to fortify the campus, and pro- and anti-Trump factions continued to clash. Meanwhile, Trump’s first administration largely spared higher education. Despite relentless criticism of universities for their putative anti-conservative bias, federal support for scientific research retained bipartisan support.

[Rose Horowitch: Colleges have no idea how to comply with Trump’s orders]

What I failed to appreciate was that the new administration was preparing the ground for a war on the American university—one that it might have carried out had the first Trump White House been better organized. In the context of crises and protests around controversial speakers, along with the growing preoccupation on campuses with offensive speech and so-called microaggressions, Trump and his allies contorted the idea of free speech to build a narrative that the university, rather than the political right, was the chief threat to the First Amendment. State after state introduced legislation, drawing on a template devised by the conservative Goldwater Institute, purportedly to defend free speech but also to enact draconian protocols for disciplining students who engaged in campus protests deemed to prevent others from speaking. (At least 23 states now have statutes in effect conferring some level of authority to state legislatures to monitor free speech on campus, demanding yearly reports, and imposing harsh new rules for student discipline.) Republican politicians began to include denunciations of universities in their talking points; in a 2021 speech, J. D. Vance declared, “We have to honestly and aggressively attack the universities in this country.”

Now the war has begun in earnest. Trump’s directives to restrict funding for science, especially the mandate to dramatically reduce National Institutes of Health grants for scientific infrastructure, equipment, and lab support—all essential components of university science—will cripple biomedical research across the country. Already, universities are reducing graduate programs and even rescinding informal offers that were made before the spending cuts were announced, and in some cases introducing hiring freezes. If the Trump administration sticks to its decision to cancel $400 million in federal grants to Columbia over the charge of tolerating anti-Semitism, we haven’t seen anything yet.

Nowhere is the assault on universities more pronounced than in the campaign to eradicate DEI. A recent Department of Education “Dear Colleague” letter warned that “using race in decisions pertaining to admissions, hiring, promotion, compensation, financial aid, scholarships, prizes, administrative support, discipline, housing, graduation ceremonies, and all other aspects of student, academic, and campus life” is prohibited. The letter purported to base its guidance on the Supreme Court’s 2023 decision striking down affirmative action, but its language went far beyond the Court’s ruling. The price of noncompliance: no federal funds. This time, I take the threat seriously.

Universities have made two general mistakes in the face of sustained right-wing criticism. First, they have behaved as if their societal value is self-evident. In fact, they need to be far more proactive in communicating the enormous contributions they make to the public good: a campaign not just to defend themselves but to remind the country that our universities are among our most crucial assets. Many of the core elements of the technologies that enable our modern lifestyle—the internet, GPS, new immunological cancer therapies, mRNA vaccines, and medical imaging, to take just a tiny sample—have emerged from academic laboratories. Whether one is concerned about democracy, how scientific research can continue to position the U.S. as a global leader, how to solve global issues such as disease and climate change, or how to maintain a competitive edge with other nations such as China and Russia, we need our universities.

[Read: A new kind of crisis for American universities]

Second, university administrators have too often assumed that because a great deal of conservative criticism of higher education has been made in bad faith, none of it is valid. The truth is that universities have not always honored their commitments to free speech, academic freedom, and open inquiry as well as they should have, and the decline in public support for universities reflects, at least in part, those failures and shortcomings.

Offices of Diversity, Equity, and Inclusion named values that for years were largely seen as benign. There was broad agreement that students from underrepresented minorities needed to have the opportunities higher education afforded but required special support to thrive in intense and often hostile academic environments for which they had little preparation or family support. Over the past decade, however, these offices grew in size and influence. With that came legitimate concerns about administrative overreach, bloat, and ineffectiveness.

At the same time, the liberal consensus was unraveling. Some faculty and students had indeed rejected the premise of free speech, noting that when power inflected all social relations, there was nothing like a level playing field; universities, they argued, should side with those lacking power and limit the speech of the powerful. Concerns about the ways in which prejudice was expressed in everyday interactions, often through unintentional slights and statements, not only surfaced as priorities for administrators but were converted into speech codes and protocols. A new language of “harm” was used to prosecute new canon wars, target faculty who offended students in the normal course of teaching, and deploy a new range of techniques to censor, punish, or “cancel” other members of the university community.

All of this came to a head in the protests after the Hamas attacks of October 7, 2023. Without any campus consensus about free speech, open inquiry, and civil discourse, an existing political impasse became even more intractable.

When, as Berkeley chancellor, I was petitioned by student and faculty groups to cancel invitations to speakers like Yiannopoulos and Coulter, I worried that to do so would be to invite censorship more broadly, and that any abrogation of free-speech rules on campus would soon be used against other political positions. I warned campus constituencies that the principle of free speech would not only protect liberals when national politics shifted—as they already had in the first Trump administration—but also help enshrine the university’s larger commitments to open inquiry and academic freedom, serious threats to which had already begun.

Now my fear that any curtailment of free-speech principles by universities would be used against universities is coming to pass. The new administration is targeting any use of race in statements or programs promoting diversity and inclusion. This effort goes far beyond admissions and hiring decisions, to the point of threatening institutions over the content of their curriculum, making a mockery of the administration’s supposed commitment to free speech. And the attacks on campus protests and DEI are just the opening salvo.

[Jonathan Chait: Anti-Semitism is just a pretext]

Governor Ron DeSantis has already signed legislation chilling instruction in disciplines including sociology and Middle East history in Florida’s public universities. Given the cuts to science funding at the federal level, we may soon see efforts to control the teaching of climate science, or biology, and maybe even evolution once again. The playbook to take “back” universities includes much more than what we have so far seen.

Federal support for scientific research, and for financial aid for students, is part of the postwar social contract that was articulated at a time when America recognized the need for as many of its citizens as possible to receive a university education and for American science to become preeminent. America’s universities, and its science, grew to be the best in the world.

This is the time to rearticulate and defend the unparalleled value of our research universities. They are the envy of nations around the globe. We attract the best and the brightest to our shores as students, researchers, and teachers. Creating these extraordinary institutions took the better part of a century, but they can be destroyed very quickly. The attack on the university may eventually backfire politically, but not before it does enormous damage. As higher-education leaders resist efforts to undermine and punish universities for their commitment to knowledge, science, and truth, they must also take care to deliver on the promises they make. Only then will the defense stand a chance of succeeding against the current assault.