The 'Big Dig' of Research Security
Can U.S. Universities Adapt to the New Rules of Internationalization?
The U.S. has developed something characteristically American in research security: a distributed, market-like network of requirements flowing through multiple federal agencies. Unlike more centralized systems in Europe and Australia, this creates a distinct dynamic where research institutions must navigate a complex web of requirements rather than working with a single coordinating body.
The 2022 National Security Presidential Memorandum 33 (NSPM-33) Implementation Guidance and 2023 National Institute of Standards and Technology (NIST) Research Security Framework seek to address emerging threats like cyber-espionage and malign foreign influence while preserving the benefits of international scientific collaboration.
The new era of "responsible internationalization"—a shift from viewing international collaboration as a means to address global challenges to seeing it primarily through an economic and national security lens—is a natural experiment in how distributed systems handle security challenges versus centralized ones.
The results might surprise us.
The Core Tension: Openness vs. Security
The core tension for the U.S. system is how a decentralized system built on a foundation of openness can retrofit security controls without breaking what made it successful in the first place.
It’s like Boston's “Big Dig” but for the entire American research enterprise — messy, expensive, and unavoidable. Similar to how the Big Dig's construction detours slowed Boston's traffic to a crawl, new security measures risk creating research bottlenecks that could persist for years.
Congress's Quiet Reshaping of Research Collaboration
Congress has discovered it can reshape research collaboration through committee letters and inquiries, without requiring new legislation.
Universities themselves are already pricing in new incentives as the cost-benefit calculations of international partnerships have fundamentally shifted. Without passing legislation, the new political consensus in Washington has raised the transaction costs of international research collaboration.
What’s Next Under Trump? Key Expectations
Is there more to come now that Trump is President? We will all find out (soon) but most observers expect:
The China Initiative (or something similar) will return. it’s revival would likely introduce new layers of scrutiny and compliance requirements for universities, particularly affecting the STEM fields where Chinese students and scholars are most concentrated.
The impact on international talent flows remains uncertain. Trump's contradictory statements and the ongoing debate within his coalition (such as the recent H1-B visa blowup on X) suggest that his actual policy implementation may differ from his campaign rhetoric. The tension between attracting skilled workers and protecting American jobs continues to shape this complex issue.
Unintended consequences are almost certain to arise.
Does friction in official research channels lead researchers to engage in different forms of "regulatory arbitrage" in scientific collaboration?
What happens to universities without elite resources? The most interesting stories won't be about Harvard or Stanford, but how cash-strapped universities manage compliance costs.
Does a nationalist-focused U.S. policy inadvertently accelerate the rise of alternative research hubs around the world? If so, where?
The Shift in Consensus: Nationalist Spirits Driving Internationalization
The most underrated aspect of this transformation is how quickly the consensus around scientific collaboration has shifted. While there is no equivalent to “responsible internationalization” in the U.S., the ethos is widely diffused across the political ecosystem, regardless of political party or who is the President.
Just as 'animal spirits' drive the markets, 'nationalist spirits' are driving the new era of internationalization.
Of AI and Economic Divergence
Like the Big Dig's massive rerouting of Boston's central artery, these two stories illustrate how America is attempting to reconstruct its research infrastructure while keeping the traffic of ideas flowing.
Story #1: The Battle Over AI Diffusion – Who Decides and What Gets Lost?
Last week, in it’s final days, the Biden administration released the "
“Trusted” partners (UK, Japan, Germany, etc.)
“Verified” partners (Brazil, Singapore), including longtime strategic allies (Israel, India)
“Restricted” entities (China, Russia, etc.)
This policy sparked polarized reactions:
On the same day, OpenAI released its "AI in America" blueprint championing a "AI compact among U.S. allies," while Anthropic's CEO, Dario Amodei, co-authored, ”Trump Can Keep America’s AI Advantage” in the Wall Street Journal that strong export controls were essential for maintaining America's AI advantage.
But there was not universal praise:
Nvidia condemned the framework as "misguided," warning it would "squander America's technological edge" and put "global progress in jeopardy." Oracle went further, arguing it will “go down as one of the most destructive to ever hit the U.S. technology industry.” Their core argument: trying to contain access through tiered restrictions is both impractical and, ultimately, counterproductive.
(We'll come back to this debate later - it reveals something crucial about the implications of the security-first “responsible internationalization” policies.)
Story #2: The Great Economic Divergence – Who Gets Left Behind?
For 70 years, America's GDP, consumption, and industrial production moved in perfect harmony (1930s-1990s). This wasn't just economics - it was the American social contract. Every part of American society participated at roughly the same pace in the nation's growth.
Then came China's WTO entry in 2001, and something broke.
What followed wasn't just an economic shift - it was a societal rupture. GDP and consumer spending kept climbing, but industrial production completely flatlined. The consequences rippled through American society: manufacturing jobs vanished, labor's share of profits collapsed, rural poverty soared, and suicide rates climbed. David Autor at MIT has documented the devastating consequences: the counties that faced the highest competition from Chinese imports were the same ones that later experienced the peak of the opioid crisis.
Out of this chaotic era, America reelected a president whose temperament matches the times—primed to shake things up with characteristic American optimism that when the dust settles, the new world order will be a better one. After decades of watching America's heartland hollow out, they view controlled chaos as preferable to continued decline.
The incoming Trump administration's response represents a radical departure from conventional economic policy. Their team isn't just making incremental changes; they're conducting the largest economic experiment since FDR, with a unprecedented mix of nine different policy potions: tariffs, deregulation, oil and gas expansion, deportations, tax cuts, cost cuts, crypto enthusiasm, medical freedom, and agency purges– an heterodox mix of policies. "The Alchemists", as Michael Cembalest characterized them, believe this cocktail of policies that can reverse two decades of industrial decline.
The Trump team is betting that the short-term disruptions excavation is part of building a superhighway for American commerce and education. However, unlike politicians of earlier eras, the new political class views universities as part of the problem. The new set of policies aims to redress perceived excesses of the past and push universities to realign with national priorities. Whether restructuring the current system can happen without damaging the very qualities that made American universities among the best in the world remains to be seen.
Three Questions Emerging from These Stories
This could usher in the most significant restructuring of international research collaborations since the Cold War era. However, unlike the binary East-West divide of that time, today’s policies must navigate a multipolar world characterized by deeply intertwined and complex research networks. Like the Middle Ages, we're entering an era where power and knowledge flow through multiple competing authorities and channels, which I’ve described as the Global Middle Ages of Internationalization.
We must ask crucial questions about power, progress, and distribution: Who decides? What gets lost? and Who gets left behind?
Governance: Who Decides?
The line between "open" and "closed" is increasingly being drawn by policymakers, not researchers or technologists. This isn't just about bureaucratic overreach - it's about who holds the authority to decide what's "necessary." The effectiveness of these policies depends on balanced and inclusive governance that incorporates diverse perspectives and avoids over-centralization. Alternative models are being overlooked in favor of top-down control.
Shifts in geopolitical alliances
New domestic legislation
Reciprocal policies abroad
Universities are already pricing in the risks of international collaboration. Some can’t afford it. Some can’t afford a PR disaster or an alumni revolt.
This results in risk aversion that narrows international research to predictable, safe research partnerships with unclear implications on what that means for scientific progress.
Innovation: What Gets Lost?
The shift toward "responsible internationalization" in academic research represents one of those classic cases where well-intentioned policy may create unintended consequences of the first order. The most concerning aspect isn't just the immediate efficiency loss, but how this reshapes the very topology of global knowledge networks.
When research becomes fragmented along national security lines, we lose the network effects that has driven many breakthrough discoveries. Consider:
Nvidia's and Oracle’s warning about export controls as “destructive” and ”squandering “America's technological edge" echo concerns from university leaders about the chilling effects of the shadow of suspicious cast over international scientific collaborations.
Similarly, universities must now weigh the payoffs of international research partnerships while pricing in downside risks of PR disasters and litigation.
The marginal cost of these restrictions grows exponentially in fields like climate science and epidemiology, where progress depends critically on dense international collaboration networks.
The real tragedy here isn't just slower progress - it's the innovations we'll never see because the right researchers couldn't connect at the right time. Markets in everything, including ideas, work best when barriers to entry are low and information flows freely.
The administrative burdens on university who will need to hire new compliance officers means that universities will forego some partnerships simply because the paperwork is too much, the approval process is too slow, or they just don’t have the staff to manage the legal complexities. It creates incentives “not to partner” and the multiplication of those incentives across thousands of universities. The result isn't just slower science - it's the breakthrough discoveries that never happen at all.
Distribution: Who Gets Left Behind?
The new security-first approach creates what amounts to technological feudalism. Middle-income countries face impossible choices:
join a technological sphere of influence at the cost of autonomy,
invest heavily in independent capabilities they might not be ready for, or
risk being left behind entirely.
The result? A dual-tiered system where resource-rich nations dominate cutting-edge research while others are relegated to peripheral roles.
What's worse, the heightened focus on security in international agreements will lead to middle-countries caught between competing geopolitical powers struggle to align with multiple, conflicting standards. This fragmented landscape hinders cooperation at precisely the moment when global challenges - from climate change to pandemics - demand greater unity.
This will create a multi-tiered system of internationalization and "intelligence divides" that represent something far more consequential than previous “digital divides”. While traditional digital divides could be bridged through infrastructure and training, intelligence divides create compounding advantages that become nearly impossible to overcome.
When one country can harness AI to accelerate progress, the gap doesn't just widen - it multiplies. As Joel Mokyr demonstrates in A Culture of Growth, once societies develop superior methods for generating and applying knowledge, these advantages tend to compound, creating permanent divergences in technological and economic capabilities.
Precision and Presumptions
Going forward, precision will be paramount. A scattershot approach risks causing irreversible damage. Treating broad categories of researchers as potential threats could backfire spectacularly. Countries in other regions—many of whose institutions are climbing global rankings—are lying in wait, ready to capitalize on any missteps by the United States.
For Universities
U.S. universities now face a classic optimization problem: how to maximize research collaboration while minimizing security risks within an increasingly intricate web of constraints. There are alternatives to "security-first" approaches that might enhance both national security and scientific progress. However, success hinges on execution, and few are giving sufficient thought to the implementation details—or the unintended consequences—of sweeping restrictions.
The implications are both operational and philosophical.
On an operational level, universities must strengthen internal compliance systems, reassess partnership agreements, and communicate new policies clearly to faculty, staff, and global collaborators.
Philosophically, universities must take active steps to preserve the fundamental values that have driven scientific progress and made U.S. research institutions global leaders.
The real question isn't binary (will universities adapt?) but rather about distribution effects of new pressures and policies on different universities and scientific fields.
For Policymakers
For policymakers, the lesson is clear: surgical precision will be far more effective than blunt force. Sustaining the dynamism of international research while safeguarding security requires:
Tailoring risk frameworks for different research domains. Climate science might warrant different protocols than quantum computing.
Using scientist-led processes to evaluate collaborations. Maintain academic autonomy while addressing security concerns.
Starting with a "presumption of openness" rather than the reverse. The burden of proof should be on those seeking restrictions, not those seeking collaboration.
There’s a huge upside to getting the policy right—it will not only accelerate science but also enhance our collective capacity to address global challenges.
The Research Security ‘Big Dig’: A Necessary Cost or Transformative Investment?
Like Boston’s “Big Dig,” America’s research security overhaul isn’t just a technical challenge—it’s a high-stakes effort to upgrade critical infrastructure for a new era while keeping the system running.
The Big Dig ultimately succeeded, delivering transformative benefits that reshaped Boston’s economy and connectivity. Yet it came at a tremendous cost, with countless unexpected complications along the way.
Similarly, research security is a complex and necessary endeavor, where the true costs—in dollars, time, lost opportunities, and institutional strain—may outstrip current estimates.
But the upside potential is immense if we get it right
What Do You Think?
What do you think? Am I missing something important here?