How Trump, Musk, and DOGE are undermining US intelligence and national security

As a long-time China scholar, and as I watch the early days of the Trump presidency, I cannot help but be reminded of the Chinese Communist Party’s assertion in the 1970s that everything Mao said or did was correct and beyond challenge. Donald Trump’s wrecking-ball approach to government reform is indeed reminiscent of Mao Zedong’s misguided attempt to achieve “better, faster, cheaper” results through ex cathedra pronouncements, disparagement of experts and bureaucratic procedures, and the dismantling the institutions of government. Mao’s denunciation of professionals and political opponents for the vague sins of “rightism” and “bureaucratism” seemingly provide the template for Trump’s crusade against the so-called radical left and “wokeism.” The Chinese Communist Party’s slavish endorsement of Mao’s assertions—that government institutions, public servants, and policies were corrupt because they did not conform to his vision for the country—resulted in prolonged and deep deterioration of government services, quality of life, and national security in China.

China’s experience should be a cautionary tale for Americans. Mao’s approach proved disastrous for the Party, the nation, and the Chinese people.

But then again, there are many distinctly American reasons for concern about the approach and actions of Trump, Elon Musk, and Musk’s young men at the Department of Government Efficiency, or DOGE (which isn’t actually a department of government). Those reasons apply across the federal government, but I want to focus on the intelligence community because of its importance to the security of our nation and the safety of Americans and American activities at home and abroad.

Identifying and assessing threats early enough to address them effectively is the work of intelligence professionals. So, too—and equally important—is the less dramatic work that they do, day in and day out, to provide information and insights to the decision makers responsible for protecting and advancing American interests. Every administration defines and prioritizes those interests differently, but all administrations need informed, tailored, timely, and objective inputs that help policymakers to cope with complexity, uncertainty, and self-serving efforts to influence their decisions. The unfettered access that Musk and DOGE have been given to agency computer systems across the government threatens to undermine US intelligence efforts in manifold ways. So do the offer of buyouts to all Central Intelligence Agency employees and other efforts to drastically reduce intelligence community staffing without anything like reasonable consideration of the likely results.

Perils of politicization. One justification given for Trump’s assault on government agencies and civil servants is the president’s declared need to remove and replace enemies embedded in what he calls “the deep state” because they lack commitment to his program. The precise definition of this commitment remains unspecified, but there has been little pushback against assertions that define it as loyalty to the president and whatever he declares to have been endorsed by his slim electoral victory.

The implication of the loyalty-above-all-else approach for the intelligence community is that objectivity and speaking truth to power must be subordinated to boosterism and confirmation of Trump-world interpretations of reality. If this happens (or is perceived to have happened), it will transform the intelligence enterprise from a trusted and unique source of information and insight into a cheerleader and propaganda outlet for the administration. Pressures to tailor collection and judgments to fit preconceptions and prove the wisdom of administration policies will encourage conformity and groupthink and make it more difficult to provide strategic warning or negative feedback.

The intelligence community differs from other entities that provide inputs to decision makers in several respects. Intelligence officials have access to classified information; the ability to utilize the incredible resources of the US government to seek and process information; individual and collective expertise, experience, and contacts inside and outside the United States; and unrivaled knowledge of the goals, preferences, preconceptions, and intelligence needs of decision makers and the missions they support. All these attributes are important, but perhaps the most important is objectivity. By design, regulation, and ethos, the intelligence community and its individual analysts are enjoined to be as objective as possible when seeking, assessing, and presenting information to reduce uncertainties in the decision-making process. Individuals, component agencies, and the intelligence community as a whole are enjoined from making policy recommendations and trained to call it as they see it, without regard to how well or badly their findings and judgments will be received by the officials they support.

Absent the perception and reality of objectivity and policy neutrality, the large, expensive, and extremely capable US intelligence community would become just another provider of ideas that has a presumed agenda of its own. That would be a fundamental and catastrophic debasement of the US intelligence advantage, making the intelligence community less useful in reducing uncertainties in the national security decision-making process, adjusting policies to suit changing conditions, gauging the efficacy of existing policies, and persuading other countries to respond positively to American warnings and suggestions. That last point bears repeating: The presumed objectivity of US intelligence community information and assessments makes it a valuable diplomatic tool. If other international actors perceive that American intelligence products and judgments are little more than propaganda for American leaders, those products and judgments will be discounted, diminishing American influence and soft power around the world.

Algorithms, expertise, and dedication. The administration’s disparagement and seeming dismissal of the expertise, experience, and commitment to mission of public servants in general and intelligence professionals in particular is worrisome, dangerous, and, again, reminiscent of Mao’s China. My experiences during more than 50 years in the intelligence community—including two stints as Assistant Secretary of State for Intelligence and Research and four years as Deputy Director of National Intelligence for Analysis and chair of the National Intelligence Council—deepened my conviction that providing world class intelligence support to American decision makers requires professionals who have deep subject matter expertise, understanding of intelligence and the intelligence process, and strong commitment to the intelligence community’s roles in the national security enterprise. The idea that any reasonably competent individual, especially one who has demonstrated fealty to the president’s program, can quickly and successfully “connect the dots” to provide timely warning, solve puzzles, and resolve uncertainties is ludicrous. It is also dangerous.

Elon Musk is obviously an intelligent man and must understand that prioritizing fealty over expertise entails significant risks. So, I assume, he must have an alternative model in mind. Though unarticulated, that model seems to propose that the way to fix deficiencies in the intelligence establishment and across the federal government is to eliminate duplication (second opinions), minimize human involvement, and allow AI algorithms a significant role in “connecting the dots” discoverable in classified and unclassified information. There are many fallacies contained in such an approach. And it is a sure-fire way to achieve an intelligence process that is, yes, smaller, cheaper, and faster—but also carries an unacceptable risk of error and provides no feedback mechanism for dissenting views.

Acquiring and applying expertise to the missions of the intelligence community requires time and commitment. Fealty to a particular president or administration; faith in the adequacy or superiority of artificial intelligence; and the loss of experienced officials through induced resignations or firings would—will—have immediate and long-term consequences. We have numerous examples of intelligence tasks being assigned to contractors who have no stake in the institutional health of the agencies in which they work and no incentive to assume the burden and responsibilities of mentorship, quality control, and long-term agility. They do their job, often very well, but within a narrow definition of what that job is. A model that depends on carpetbaggers to perform basic intelligence functions will not perform at levels needed to meet the requirements and expectations of the US national security enterprise.

Data security. Politicians and pundits have raised many appropriate concerns about the access to data being given to largely unvetted and unknown DOGE programmers. If it was problematic that Chinese hackers gained entry to databases containing the personal information of US government employees and ordinary citizens, why is it less worrisome that Musk’s largely unknown geeks have access to the same information—and more? The reasons for concern are more numerous and serious for the intelligence community than for most other government agencies because, presumably, when DOGE accesses intelligence agency databases, it will have access to even highly classified intelligence reports.

Those who argue that DOGE operatives will not delve into sensitive intelligence reporting because they need only administrative access ignore an important question: How do we know what access they will have, given that virtually nothing about them or their mandate has been made public? Even more worrisome, how do we know that the team does not include an Edward Snowden? Snowden was a contract employee with system administrator access to intelligence community computer systems who stole and leaked thousands of highly classified documents in 2013. I have absolutely no information suggesting that one or more members of the Musk team are imitating Snowden’s collection of materials for nefarious purposes. But neither do I have reasons for confidence that this will not or could not happen.

The concern of Americans about who has access to their personal and professional data is warranted; more should be done to address this concern. But, once again, intelligence is a special case. The intelligence community holds millions of digitized documents, many of them highly classified. How information was obtained is often the reason for classification; that reason is called “protection of sources and methods” in intelligence service jargon. That’s to say, the content of a report or reports may be less sensitive than the means used to obtain it.

Protecting technical capabilities and the people who provided information to the United States or to one of our allies or partners who share intelligence with us is vital to US national security. The world is a big place, and US interests are global. It is extremely helpful—essential—to maintain intelligence-sharing arrangements with other countries. If allies worry that we are unwilling or unable to protect their information and how it was obtained, they won’t share it—or they will dumb it down in ways that makes it less useful or even useless to us.

In the first months of Mao’s Cultural Revolution, one of the questions on an exam in one of my undergraduate classes asked what Mao hoped to accomplish in his assault on the Communist Party, the government, and China’s public servants, and whether he would succeed. My answer to the second part of that question was that the campaign would be extremely disruptive but, in the longer run, it could not succeed. I have the same judgment about what Trump and Musk are doing.

One need not spend much time in Washington to recognize that redundant or wasteful federal programs and activities are approved by Congress; they are not initiated by civil servants of the executive branch. The place to start looking for inefficiencies in government, then, is in the authorization and appropriation process, not the files of executive agencies. Tax breaks, regulatory relief, or other benefits to favored business sectors and/or as rewards for major campaign contributors are indeed sources of enormous inefficiency. Some unwarranted government benefits endure for years without scrutiny. There are certainly expenditures that could safely be reduced or eliminated entirely, in most and perhaps all executive agencies.

But if one is serious about reducing costs and improving efficiency of the federal government, it would be far better and smarter to start with the authorization and appropriation process than to plow wholesale through databases containing vast amounts of extremely sensitive information, along the way attempting to rid the government of experienced intelligence experts who help keep the United States secure, regardless of the party or proclivities of any particular president.

Editor’s note: The views expressed in this publication are the author’s and do not imply endorsement by the Office of the Director of National Intelligence, the intelligence community or any US government agency.