Six months after a hantavirus outbreak threatened to go international, and with an Ebola emergency still stubbornly refusing to wrap itself up, comes word that a Russian laboratory worker may have died from pneumonic plague. Nothing says "global health community" quite like a deadly pathogen that respects no borders while the humans tasked with containing it remain locked in geopolitical squabbling, mutual suspicion, and a shared commitment to not returning phone calls.
Russian officials have downplayed the reports, and the risk of international spread appears low. Untreated pneumonic plague is often fatal, yes, but early antibiotics and localized ring containment reliably curb outbreaks. So the good news: this is probably not the sequel. The bad news: the episode exposes the world's biosafety and biosecurity defenses as a patchwork of hope and duct tape. We face zoonotic spillovers of novel pathogens, dangerous gain-of-function research (that's the study of how pathogens become more deadly or transmissible, for those keeping score at home), and weak laboratory safeguards - all now amplified by advances in artificial intelligence. Because what this timeline needed was AI that could help malefactors enhance pathogenicity or cook up new pathogens entirely.
The suspected plague case in Russia is eerily reminiscent of early COVID-19 and SARS responses. The Kremlin failed to report the case to the World Health Organization, as legally required under international health regulations. Instead, the WHO learned about it through media reports - the international public health equivalent of finding out you've been fired via a tweet. When the WHO officially requested verification, Russia offered vague assurances. There is no transparency regarding the research in the Siberian lab or the circumstances of infection.
The Russian lab worker's exposure to pneumonic plague raises serious biosecurity questions. So far, the Irkutsk facility appears to focus on civilian surveillance and research on regionally endemic diseases. Pathogen research on bacteria like Yersinia pestis (plague) or anthrax can be perfectly legitimate - developing better antibiotics, for instance. But it can also be exploited for less noble ends, which is the diplomatic way of saying "we have no idea what's happening in there."
Russia and the U.S. hold the only known official stocks of the smallpox virus, which would be catastrophic in the wrong hands. U.S. intelligence agencies, meanwhile, assess that Russia maintains an active, offensive biological weapons program. So that's reassuring.
Without appropriate biosafety controls, hazardous research risks exposing lab workers and causing leaks that could trigger major outbreaks. Currently, there are no international systems aimed at regulating or inspecting facilities that undertake gain-of-function research, establishing mandatory biosafety standards for labs, or monitoring AI capabilities that could enhance and weaponize existing pathogens or create new ones. The need for better global biosafety and biosecurity rules is, to use the technical term, screamingly obvious.
The state of biosecurity worldwide has always been fragile. Unlike its counterpart, the Chemical Weapons Convention, the Biological Weapons Convention is brief, weak, and lacks a robust system of verification and enforcement. It is likely that 50 or more governments actively support dual-use research allowed under the treaty that could be adapted for prohibited purposes. Even purely civilian biological research facilities often lack core security controls or remain vulnerable to malicious actors. With rapid, poorly monitored advances in AI-driven protein manipulation and mirror-life research, the possibilities for accidental or intentional misuse are growing exponentially. Anthropic recently disclosed five separate incidents of Claude being used in research linked to biological weapons development, but disclosures across major AI developers make clear that providers often fail to detect misuse for months - which is a bit like a security guard who notices the break-in during the cleanup.
The WHO issues guidance to minimize biosecurity risks, but even that framework is rapidly becoming obsolete. The U.S. National Institutes of Health recently released an updated Draft Biosafety Policy for Research Involving Biohazards, which attempts to move research institutions toward a risk-based oversight framework. Yet even within that document, the threats posed by artificial intelligence remain undefined and addressed only indirectly - the policy equivalent of gesturing vaguely at the elephant in the room.
So what should be done? First, it is time for the Biological Weapons Convention to adopt a Chemical Weapons Convention-like notification and verification regime. Under that treaty, governments must provide a verifiable list of any past or present weapons programs, storage facilities, production sites and relevant industry activities, as well as a risk-based classification system. International inspectors are authorized to verify those declarations under a risk-weighted system where higher-risk facilities are inspected more frequently. Revolutionary concept: actually checking.
Second, there is already a precedent for international verification that could be expanded without amending the Biological Weapons Convention. Historically, specialized WHO inspection teams have verified smallpox research facilities in the U.S. and Russia. Targeted international cooperation through the United Nations or the Chemical Weapons Convention framework could perform a similar inspection function.
Third, significant funding is required simply to bring laboratories across much of the world up to existing WHO guidelines and best practices. Many countries, including the U.S., directly funded these biosafety investments before severe budget cuts disrupted programming. This funding and associated capacity-building programs must be restored.
Finally, the WHO Pandemic Agreement currently languishes as countries remain at an impasse over how to govern pathogens with pandemic potential. Key provisions of this agreement target biosafety and biosecurity. Finalizing the WHO Pandemic Agreement is an indispensable complement to broader global security measures.
The time for coordinated international standards and compliance mechanisms for biosafety and biosecurity - especially given rapid AI advances - is now. Governments must move past the current impasse to protect the rapidly expanding life sciences field and create a safer world. Or they can keep doing what they're doing and hope for the best. History suggests the second option has a body count.
Sam Halabi is the director of the Center for Transformational Health Law at Georgetown University's O'Neill Institute for National and Global Health Law and a professor at Georgetown's School of Health. He is also an affiliate researcher at the school's Center for Global Health Science and Security.
Lawrence O. Gostin is a distinguished university professor, the co-faculty director of the O'Neill Institute and the founding O'Neill Chair in Global Health Law. He is director of the WHO Collaborating Center on Global Health Law.