Biosecurity and Dual-Use Research

The Select Agent Program

The foundational US biosecurity framework is the Select Agent Program, jointly administered by the CDC and USDA APHIS.

The program designates certain pathogens and toxins as Select Agents — biological agents with the potential to pose severe threats to public health or agriculture. Entities possessing, using, or transferring Select Agents must:

• Register with the program
• Conduct security risk assessments of personnel
• Implement specific biosafety and biosecurity measures
• Report incidents (releases, thefts, loss of inventory)
• Submit to FBI background checks for all personnel with access

The list of Select Agents includes:

• HHS Select Agents — Pathogens primarily affecting humans (Ebola, Marburg, plague, smallpox, anthrax, certain neurotoxins)
• USDA Select Agents — Pathogens primarily affecting animals or plants (foot-and-mouth disease virus, rinderpest, certain plant pathogens)
• Overlap Agents — Pathogens with both human and animal/plant impact

The program emerged in its current form after the 2001 anthrax letters and was strengthened by the Public Health Security and Bioterrorism Preparedness and Response Act of 2002. It applies to all entities possessing Select Agents — universities, government labs, private companies, hospitals.

The Select Agent program is one of the more aggressive biosecurity frameworks globally. Critics argue it's burdensome (researchers report substantial time investments in compliance), arbitrary (the list of designated agents reflects historical risks more than current threats), and incomplete (it covers possession but says little about emerging synthetic biology techniques). Supporters argue it's essential infrastructure for preventing dangerous pathogens from being stolen, misused, or accidentally released.

The program has prevented some incidents. It has not prevented all — there have been multiple reported lapses, including a 2014 incident where the CDC itself accidentally exposed dozens of staff to potentially active anthrax, and the 2015 discovery that the US Army had been inadvertently shipping live anthrax to dozens of labs over a decade. The program is necessary but not sufficient.

---

Gain-of-Function Research and the Pause

In 2011, two research teams — one led by Ron Fouchier in the Netherlands, one by Yoshihiro Kawaoka at the University of Wisconsin — independently created versions of the H5N1 avian influenza virus that could spread between mammals (specifically, ferrets) through airborne transmission. The natural H5N1 was highly lethal in humans but had not yet evolved to spread efficiently between people. The researchers' modifications gave it that capability.

These studies were intended to help understand pandemic potential and inform vaccine development. They also produced what some experts called "potential pandemic pathogens" — engineered viruses that could, if released or replicated, cause mass casualties.

The research triggered intense controversy. The National Science Advisory Board for Biosecurity (NSABB) initially recommended that the methods sections of the papers be redacted before publication. The recommendation was later modified, and the full papers were published.

In 2014, the US government imposed a funding pause on certain types of gain-of-function research — specifically, work that "may be reasonably anticipated to confer attributes to influenza, MERS, or SARS viruses such that the resulting virus has enhanced pathogenicity and/or transmissibility in mammals."

The pause was controversial. Some scientists argued it shut down important pandemic preparedness research. Others argued the pause didn't go far enough. Critics noted that some excluded research continued anyway through workarounds or in international collaborators not subject to US restrictions.

In 2017, the pause was lifted and replaced with the HHS Potential Pandemic Pathogen Care and Oversight (P3CO) Framework. Under P3CO, federally funded research that meets specific criteria for creating pathogens of pandemic potential undergoes additional review by an HHS-level committee.

The P3CO framework has been criticized from multiple angles. Some research that should arguably have triggered review (notably, certain SARS-related coronavirus work) was apparently not flagged. The committee's deliberations are not transparent. The criteria are vague enough to leave significant interpretive room.

The COVID-19 pandemic intensified these debates dramatically. Whether SARS-CoV-2 emerged through natural zoonotic spillover or through a laboratory incident remains formally unresolved, though most virologists favor natural origin. Either way, the question of whether existing biosecurity policy adequately addresses the risks of coronavirus research has become a central biosecurity policy debate. Multiple congressional hearings, an HHS Office of Inspector General review, and academic reviews have all examined whether US biosecurity policy is fit for purpose. The reforms continue to be developed.

---

Dual-Use Research of Concern (DURC) Policy

Beyond specific funding pauses, the broader policy framework for Dual-Use Research of Concern governs research that has high biomedical value but could also be misused.

The current policy structure has evolved substantially over the past decade:

The 2012 United States Government Policy for Oversight of Life Sciences DURC. Required federal agencies to identify DURC research and take steps to mitigate risks. Defined DURC as research with seven specific pathogens that could be expected to enable: enhanced harmful properties; disruption of beneficial effects of therapeutics; evasion of detection; enhancement of stability or transmissibility; alteration of host range; production of harmful effects; or weaponization.

The 2014 Institutional DURC Policy. Extended DURC oversight to institutions (universities, research institutes) conducting federally funded research.

The 2024 Update. In May 2024, the White House Office of Science and Technology Policy released a major update to DURC policy, combined with new policy on Potentially Pandemic Pathogens (PEPP). Key changes:

• Expanded list of agents covered
• New category of "experiments of concern" — specific types of research requiring enhanced oversight regardless of agent used
• Specific provisions for PEPP research
• Clearer institutional responsibilities and reporting requirements
• Extension to non-federally funded research (in some categories)

The 2024 update is the most significant revision of US biosecurity policy in over a decade. It addresses several long-standing criticisms but also faces criticism for: (1) still relying primarily on self-identification by researchers, (2) limited transparency in oversight decisions, (3) inadequate attention to emerging technologies (especially AI-enabled biotech), and (4) gaps in coverage of private and international research.

The fundamental challenge: DURC policy is essentially a system for asking researchers and institutions to flag their own potentially dangerous research. The incentives for flagging are weak (delays, restrictions, reputational risk); the incentives for not flagging are strong (continued funding, faster publication, no oversight). Whether this system can reliably identify and govern the most concerning research is an open question.

---

AI, Benchtop Synthesis, and the New Biosecurity Landscape

Traditional biosecurity policy was built for a world where dangerous biotech required substantial resources, specialized facilities, and decades of training. That world is changing fast.

Benchtop DNA synthesis. Commercial DNA synthesis services can now produce custom DNA sequences and ship them to researchers. The industry has implemented voluntary screening — companies check ordered sequences against databases of dangerous pathogens to prevent (for example) someone ordering the smallpox genome piece by piece. But the screening is voluntary, the databases are incomplete, and bench-top DNA synthesizers — devices that can produce DNA in-house without going through a commercial provider — are becoming smaller and cheaper.

The International Gene Synthesis Consortium has established voluntary screening standards. Most major commercial DNA synthesis providers participate. But the gaps are real, and the long-term trajectory is toward more accessible, less regulated synthesis capacity.

AI and protein design. AI tools like AlphaFold (for protein structure prediction) and various generative AI systems for protein and DNA design dramatically reduce the expertise required to develop biological products. Tasks that required PhD-level expertise five years ago are now accessible to anyone with computational resources. This is mostly good — it democratizes biotech. It also expands the population of people who could, in principle, develop dangerous biological products.

Several recent studies have demonstrated that current large language models can provide substantial help to a non-expert trying to develop a biological weapon, though the results are contested. A 2024 RAND report concluded that current AI provides only marginal uplift, while other analyses argue the risk is more significant. The field is genuinely uncertain.

The 2023 Executive Order on AI. EO 14110 (October 2023) included specific provisions for biosecurity, requiring federal agencies to develop guidelines for AI tools that could be used to create biological threats and requiring DNA synthesis screening as a condition for federal funding.

Emerging policy proposals:

• Mandatory (rather than voluntary) DNA synthesis screening
• AI model evaluations specifically for biosecurity risks
• Restrictions on releasing certain AI model weights publicly
• Expansion of DURC framework to AI-enabled biotech
• Coordination with international partners on biosecurity (the Bletchley Park AI Safety declarations from 2023 included some biosecurity language)

These proposals are contested. Some experts argue the risks are overstated and the proposed restrictions would damage legitimate research. Others argue current frameworks are dangerously inadequate. The debate is unresolved.

What's clear: biosecurity policy designed for the 2000s does not fit the technological landscape of the 2020s. The 2024 DURC update is one response. More will follow. This is the policy area where the pacing problem — the gap between technology and governance — is most acute.

---

Wait, Actually...

The most successful biosecurity intervention in human history happened through international cooperation that's almost impossible to replicate today.

Smallpox eradication was achieved through a WHO-led campaign that ended natural smallpox transmission by 1977. After that, only two officially sanctioned stocks of variola virus remained — at the CDC in Atlanta and the State Research Center of Virology and Biotechnology (VECTOR) in Koltsovo, Russia.

For decades, debates have raged about whether to destroy the remaining stocks. The WHO has repeatedly considered destruction, and each time the decision has been deferred. The arguments for keeping the stocks: defensive research, vaccine development. The arguments for destroying them: any retained stocks pose catastrophic risk if released, and the genetic sequence is now publicly available anyway, meaning the virus could theoretically be synthesized from scratch (the 1918 flu reconstruction proved this is possible).

In 2024, the question of smallpox stock destruction is still unresolved. But the deeper point is this: smallpox eradication required Cold War-era US-Soviet cooperation that's nearly unimaginable today. Russian and American virologists worked side by side in remote villages, with shared resources and protocols, to eliminate a disease that had killed an estimated 300 million people in the 20th century alone.

We can't have that level of biosecurity cooperation now. The Wuhan Institute of Virology controversy, Russian biological weapons programs, accusations and counter-accusations between major powers — the trust required for genuine biosecurity collaboration has eroded in ways that may be hard to recover. Modern biosecurity policy operates against a backdrop of geopolitical tension that the smallpox era didn't face.

The biggest unsolved problem in biosecurity isn't technical. It's diplomatic.

---