Bioweapons Governance: What Each Position Is Protecting

On October 4, 2001 — twenty-three days after the attacks on the World Trade Center and the Pentagon — letters containing anthrax spores arrived at the offices of two United States senators and several media organizations. By the time the attack was contained, five people were dead, seventeen more infected, and tens of thousands had received antibiotic prophylaxis. The Hart Senate Office Building was closed for three months. The FBI investigation, the largest in the Bureau's history, concluded seven years later that the letters had been sent by Bruce Ivins, a microbiologist at the United States Army Medical Research Institute of Infectious Diseases at Fort Detrick — a federal biodefense facility. Ivins died by suicide as prosecutors were preparing charges. The case illustrated something that the entire bioweapons policy community had known but struggled to convey: that the threat of biological weapons does not come only from states with large programs. It can come from a single person with institutional access and a grievance.

The international community had, in 1972, adopted what remains one of the most sweeping arms control agreements ever negotiated. The Biological Weapons Convention prohibits the development, production, stockpiling, and transfer of biological agents for hostile purposes — not just their use, but their existence as weapons. By 2026, 183 states have joined it. The convention is, in the terms used by arms control lawyers, categorical: unlike the Nuclear Non-Proliferation Treaty, which explicitly legitimizes existing arsenals for five states, the BWC treats biological weapons as simply impermissible, for everyone. It has no verification mechanism. It has no independent inspectorate. Its annual budget is approximately four million dollars — less than the operating budget of most medium-sized police departments. When the United States rejected a draft verification protocol in 2001, citing concerns about industrial trade secrets and intelligence methods, the decade-long effort to build an oversight body collapsed, and has not been seriously revived.

The people who think seriously about bioweapons governance do not agree about much. What they agree on is that the problem is getting harder: synthetic biology has lowered the technical barriers to engineering dangerous pathogens; the global life sciences workforce is larger than it has ever been; and the information hazards created by legitimate scientific research — the horsepox reconstruction, the H5N1 gain-of-function experiments, the published synthesis routes for dangerous pathogens — cannot be unpublished. The disagreements are about what to do about it.

What BWC and arms control advocates are protecting

The categorical norm against biological weapons — and the argument that its strength comes precisely from its categorical character, which must be defended actively rather than assumed to be self-sustaining. The 1972 Biological Weapons Convention is, in one reading, a remarkable achievement: the first arms control treaty to prohibit an entire class of weapons before they had been used at scale in a major war, adopted with genuine multilateral consensus including both Cold War superpowers. Filippa Lentzos at King's College London, who has studied the BWC for two decades, argues that the convention's prohibition on development and production — not just use — makes it fundamentally stronger than the 1925 Geneva Protocol, which banned use but permitted stockpiling. Arms control advocates are protecting the norm itself: the understanding, shared by the overwhelming majority of states, that biological weapons are not a legitimate instrument of military power. They argue that norms do not persist automatically; they require institutional reinforcement, universal membership, and visible accountability. A convention that everyone signs and no one verifies is a convention that powerful actors can defect from without immediate consequences — and that smaller states have reason to hedge against. The absence of a verification mechanism is, in this view, not a manageable technical gap. It is a fundamental vulnerability in the architecture of the prohibition.

The project of closing the verification gap — and the recognition that the 2001 collapse of verification negotiations was a political choice that could, in principle, be reversed. The draft verification protocol that the United States rejected in 2001 had been under negotiation since 1995. It would have established a technical secretariat, mandated declarations of relevant facilities, and enabled both routine and challenge inspections — a structure roughly analogous to what the Chemical Weapons Convention has in the Organisation for the Prohibition of Chemical Weapons. The US objections were not frivolous: dual-use biological facilities are genuinely harder to inspect than chemical plants, the risk of industrial espionage under the cover of inspection is real, and the protocol's proponents acknowledged that it would not reliably detect a sophisticated state program. Arms control advocates argue that the rejection was nonetheless a mistake: an imperfect verification mechanism is better than no mechanism, because it raises the cost of defection, generates evidence when defection occurs, and communicates to the world that the major powers take the convention's obligations seriously. Gregory Koblentz at George Mason University, in his study of the BWC's structural weaknesses, argues that the lack of verification has made it easier, not harder, for states with ambiguous programs to avoid accountability — and that the dominant global power's rejection of oversight for its own facilities has undermined its authority to demand transparency from others.

What national biodefense advocates are protecting

Operational readiness against a threat that documented history confirms is real — and the recognition that effective biodefense requires understanding what offensive programs actually look like. Ken Alibek — Kanatjan Alibekov — defected from the Soviet Union in 1992 after serving as deputy director of Biopreparat, the Soviet bioweapons program that continued to operate throughout the BWC period. In his 1999 memoir Biohazard, Alibek described a program that at its peak employed sixty thousand scientists, engineers, and production workers, weaponized plague, smallpox, and Marburg hemorrhagic fever, and developed anthrax strains resistant to the vaccines held by Western biodefense programs. The Soviet Union had signed the Biological Weapons Convention in 1972. The program was one of the most consequential arms control violations of the twentieth century and was never publicly acknowledged by the Russian government. Biodefense advocates are protecting against the naive assumption that the existence of the BWC means the threat does not exist. They argue that intelligence services, military medical research, and consequence management capabilities are necessary precisely because some actors have defected and will defect from the prohibition — and that a nation that cannot detect a biological attack, identify its agent, or develop countermeasures rapidly is not protected by a treaty. It is undefended.

The medical countermeasure development pipeline — and the argument that the line between offensive and defensive biological research, while conceptually important, is not always operationally clear. Developing a vaccine against a weaponized pathogen requires understanding how the pathogen was modified. Developing a diagnostic for an engineered agent requires knowing what modifications to look for. The institutional embodiment of this tension is USAMRIID — the same facility whose researcher sent the 2001 anthrax letters, and also the facility that produced the anthrax vaccine and the early work on Ebola countermeasures. Biodefense advocates argue that this is not hypocrisy but necessity: the BWC's permissive carveout for "prophylactic, protective, or other peaceful purposes" is not a loophole but an acknowledgment that effective defense requires research that, in a different institutional context, would constitute weapons development. The question they are navigating is not whether to have a biodefense program — they argue you must — but how to maintain the credibility of the defensive characterization when the research itself is dual-use. Richard Danzig, former Secretary of the Navy and a longtime student of catastrophic biological risk, has argued that the United States government's approach to biodefense has historically been more coherent on stockpiling countermeasures than on building the surge capacity to deploy them — and that the actual protection offered is substantially lower than the program's budget would suggest.

What dual-use research governance advocates are protecting

The integrity of the scientific enterprise — specifically, the possibility of making deliberate governance decisions about which research questions should be pursued, before those decisions are made by default through the accumulation of published results. In 2011, two research groups — Ron Fouchier's laboratory at Erasmus Medical Center in Rotterdam and Yoshihiro Kawaoka's laboratory at the University of Wisconsin — independently reported that they had created influenza H5N1 strains capable of airborne transmission between ferrets. H5N1 had previously required direct contact for transmission between mammals; in the rare cases where it infected humans, it killed roughly 60 percent of them. The papers described, in scientific detail, the specific mutations that enabled airborne transmission. The United States National Science Advisory Board for Biosecurity initially recommended that the journals not publish the methods sections — a recommendation that was reversed after further deliberation and that generated some of the most contentious debates in recent scientific history. Marc Lipsitch at Harvard and Thomas Inglesby at Johns Hopkins argued in a widely cited piece that the research represented a category of work where the expected costs — the probability of pandemic escape multiplied by the consequences — exceeded the expected benefits, regardless of what was published. They were arguing not against science but for a governance process capable of making that calculation before the experiment was run, rather than after the paper was ready.

The possibility of governing information hazards before the information has spread — which requires building governance structures with real authority before they are needed in a crisis, not designing them in the wake of one. The 2014 federal moratorium on certain gain-of-function research and its 2017 replacement by a P3CO (Potential Pandemic Pathogen Care and Oversight) review framework represent attempts to build such governance. Dual-use research advocates argue that these mechanisms are directionally correct but structurally weak: the review process is advisory, the definitions of what triggers review are contested, and the institutional pressure within the life sciences community to publish and to fund ambitious research works consistently against precaution. Richard Ebright at Rutgers, one of the most persistent critics of gain-of-function governance, has argued that the P3CO framework has effectively exempted the research it was designed to review by defining its scope narrowly. What this position protects is the space for governance to be genuine rather than performative — for the bodies charged with oversight to have the authority, independence, and scientific competence to say no when the risk calculus warrants it, rather than ratifying decisions that funding agencies and researchers have already made.

What global health equity advocates are protecting

The recognition that natural pathogens are the dominant biological threat to human life by orders of magnitude — and that a governance frame organized around bioterrorism and state weapons programs systematically misdirects resources and attention. COVID-19 killed at least seven million people in its first three years by official counts; excess mortality analyses suggest the true figure is several times higher. Seasonal influenza kills three hundred thousand to six hundred fifty thousand people annually. Tuberculosis kills approximately one and a half million people per year, almost all of them in low- and middle-income countries. Malaria kills six hundred thousand to eight hundred thousand, the majority children under five. The total death toll from biological weapons attacks in the twentieth century — including the Japanese Unit 731 program, the only large-scale use of biological weapons in warfare — is estimated at roughly ten thousand people. Global health equity advocates are not dismissing the bioweapons threat; they are insisting that the framing of biological risk around exotic weapons rather than endemic pathogens produces a policy environment in which biosafety level four laboratories in wealthy countries receive substantial investment while basic health systems in low-income countries remain under-resourced. Peter Hotez at Baylor College of Medicine has argued that the most effective biosecurity investment the United States could make would be funding vaccine development for neglected tropical diseases — not because bioterrorism is unreal but because the pathogens that most reliably kill the most people are the ones currently given the least governance attention.

Equitable access to medical countermeasures — and the argument that a global biosecurity architecture that cannot distribute vaccines, diagnostics, and treatments equitably in a crisis is not a biosecurity architecture at all. The COVID-19 pandemic's vaccine rollout made the access problem explicit in a way that previous public health crises had not. High-income countries secured advance purchase agreements for the majority of global vaccine supply; low-income countries waited. COVAX — the international mechanism designed to ensure equitable access — fell short of its coverage targets in its first year partly because wealthy nations had already purchased more doses than their populations required. The TRIPS waiver debate — over whether intellectual property rules should be suspended to allow generic production of vaccines and therapeutics in countries without domestic manufacturing capacity — exposed the same structural inequity that equity advocates had been describing for years: the governance architecture for biological threats is designed primarily by, and primarily serves, states with the resources to build their own capabilities. The Global Health Security Index, published by Johns Hopkins and the Nuclear Threat Initiative, ranks countries on biosecurity readiness; in its 2021 edition, no country scored above 75.9 out of 100, and the median score among low-income countries was approximately 19. Global health equity advocates are protecting the argument that security from biological threats requires a functional global public health system — and that a governance architecture that treats low-income countries as sites of threat rather than partners in protection will fail to address the conditions under which the next pandemic will emerge.

What the argument is actually about

Whether the categorical prohibition on biological weapons can be maintained as the underlying technology becomes cheaper, more accessible, and more dual-use — or whether the prohibition requires active institutional reinforcement that the current framework does not provide. The BWC's strength is its normative force: the overwhelming consensus among states that biological weapons are illegitimate. Its weakness is that normative force without verification is difficult to sustain against actors — state or non-state — who calculate that defection is unlikely to be detected or punished. The question that arms control advocates and biodefense advocates are actually arguing about is whether that normative force can be preserved through treaty reinforcement and verification, or whether it requires maintaining a credible deterrent capacity — and whether those two strategies are compatible or contradictory. A state that invests heavily in biodefense research, citing the Soviet precedent, generates the same dual-use concern from its adversaries that justifies its own investment. The resulting dynamic is familiar from nuclear arms control: each party's defensive posture looks offensive to the other, and the mutual escalation of precautionary capacity creates the threat environment that made the precaution seem necessary.

Whether the distinction between offensive and defensive biological research is coherent enough to bear the weight the governance architecture places on it. The BWC's carveout for "prophylactic, protective, or other peaceful purposes" was understood, when the treaty was negotiated, as a relatively clear distinction: developing a vaccine is different from weaponizing a pathogen. Synthetic biology has made the distinction harder to maintain. Enhancing a pathogen to test vaccine efficacy requires making the pathogen more dangerous. Developing diagnostics for engineered agents requires knowing what engineering looks like. Building consequence management capacity requires doing the simulations and producing the materials that would be used in an attack. Dual-use research governance advocates argue that this blurring does not make the distinction meaningless — it makes it harder to apply mechanically, and more dependent on the institutional context and intent of the research. What they are protecting is a governance process capable of making that determination with real authority, rather than leaving it to grant committees and scientific journals after the research has been completed.

Whether biosecurity is primarily a military and intelligence problem or primarily a public health problem — and whether the framing determines the governance architecture that gets built. This is the question global health equity advocates place at the center, and it is genuinely consequential. A biosecurity frame organized around intentional attacks, state programs, and arms control produces one set of institutions: intelligence services, military biodefense laboratories, classified assessments, national security councils. A biosecurity frame organized around biological threats of all origins — including the natural emergence and spread of novel pathogens — produces a different set: WHO, national public health agencies, pandemic preparedness infrastructure, global surveillance networks. The COVID-19 pandemic demonstrated that the second frame had been systematically underinvested relative to the first, and that the consequences of that underinvestment were severe. It did not demonstrate that the first frame was unnecessary — the documented history of state bioweapons programs and the growing accessibility of the relevant technologies make that argument difficult to sustain. What it demonstrated is that the two frames are not alternatives. A governance architecture adequate to the full spectrum of biological risk requires both — and resources, authority, and institutional coherence adequate to the challenge. The current architecture has neither.

Bioweapons governance sits at an uncomfortable intersection: the weapon that arms control advocates most want to prohibit absolutely is the weapon that public health advocates insist is less dangerous than the pathogens that emerge naturally every generation. Both are right. A pandemic engineered and released deliberately would be catastrophic; a pandemic that emerges naturally has already been catastrophic, and the infrastructure built to prevent it is the same infrastructure needed to limit the damage from the deliberate version. The debates about BWC verification, gain-of-function governance, and global health equity are not separate conversations happening in separate policy spaces. They are the same conversation about whether humanity can govern biological risk before the next demonstration of why it must.