Portside

Some of the world’s deadliest toxins are found in marine creatures such as the puffer fish and the blue-ringed octopus. For many, there is no antidote. So when U.S. Department of Defense (DOD) officials in 2019 confronted their Chinese counterparts with concerns about experiments with marine neurotoxins being conducted in China’s military labs, they were hoping for reassurance. Prior discussions had included “good dialogue” on how to keep such research from taking a nefarious turn, says a former DOD official involved in the sensitive talks.

But the neurotoxin queries hit a nerve. “We got completely slammed with disinformation,” including assertions that the United States was operating its own bioweapons programs, the former official says. China broke off talks, and then in 2020 Chinese officials alleged that the U.S. Army had released the COVID-19 virus in Asia as a bioweapon—a charge the U.S. vehemently denied.

The bitter breakdown highlighted what many see as enduring weaknesses in the Biological Weapons Convention (BWC), a landmark arms control treaty that this year marks its 50th anniversary. Architects of the BWC, which has been ratified by China, the U.S., and 186 other countries, hoped it would eliminate weapons “repugnant to the conscience of mankind.” But the convention will need much stronger teeth if it is to keep humanity safe, said researchers who gathered last week at a meeting in Washington, D.C., sponsored by the U.S. National Academies of Sciences, Engineering and Medicine to review the convention’s first 5 decades.

One major flaw, speakers said, is that the BWC lacks a mechanism to verify compliance—or even meaningful transparency measures. If international monitors can’t gain access to labs, or glean clues to what is happening there, it’s hard to judge when dual-use research—peaceful science that can have military applications—crosses a line. Adding to the urgency are rapid advances in the life sciences including synthetic biology and gene editing, enhanced by artificial intelligence (AI). They could lead to “weapons that are nastier than what’s found in nature—more transmissible and more deadly. Or resistant to existing vaccines or drugs,” says Jaime Yassif, who leads the biosecurity program at the nonprofit Nuclear Threat Initiative (NTI).

BWC working groups and nonprofits have been exploring a variety of approaches to strengthening the pact. Monitors could use AI to sniff out suspicious signals in trade data, scientific articles, and satellite imagery, for example, while companies that sell DNA sequences could use sophisticated screens to weed out those that might be used to construct weapons. Drug companies could be coaxed to stand in as proving grounds for inspections should treaty parties adopt site visits as a way to force violators to move or shut down work.

But efforts to strengthen the BWC face stiff headwinds. In December 2024, a treaty meeting ended in acrimony after Russia derailed consensus on measures for international cooperation and assistance in implementing the BWC and for the provision of science advice to treaty members. In his closing statement, Italy’s top negotiator, Leonardo Bencini, lamented the impasse. “If a biological weapon were to be used tomorrow, we would be caught completely unprepared. The world would look at us and would ask: ‘What have you done to prevent this?’” 

The first international agreement to ban an entire category of weapons of mass destruction, the BWC has led to the closure of more than 20 offensive bioweapons programs since it entered into force on 26 March 1975, including the clandestine weaponization of everything from deadly human pathogens such as anthrax and smallpox to rinderpest, a virus lethal to cattle, and stem rust fungi against wheat and rye. “It’s important to remember that before the BWC, the U.S. had one of the largest biological weapons programs in the world. Leaders chose to walk away from that capability,” says Gigi Kwik Gronvall, a biosecurity expert at the Johns Hopkins University Center for Health Security.

Still, although bioweapons are unwieldy—they can be tricky to deploy and sicken the aggressor’s own troops or civilians—they remain appealing because they are cheap and terrifying. In an April 2024 nonproliferation report, the U.S. Department of State decried what it views as worrisome dual-use research in China and Iran. And it concluded that two nations—North Korea and Russia—currently have programs to develop and stockpile offensive bioweapons. Unclassified details are sketchy, but “there’s no uncertainty in that assessment,” says Andrew Weber, a senior fellow at the Council on Strategic Risks who led threat reduction programs for the U.S. government, including dismantling a Soviet anthrax facility in Kazakhstan in the 1990s. The public report, he says, is “the visible tip of an underwater mountain of intelligence.”

According to the report, Russia has spent millions of dollars renovating Soviet-era bioweapons labs of its 48th Central Scientific Research Institute. In October 2024, The Washington Post highlighted satellite images of the ongoing refurbishment at Sergiyev Posad-6, the institute’s virus lab near Moscow. Before the Soviet Union collapsed, researchers there weaponized smallpox, Ebola, and other pathogens. The new imagery indicates a “massive buildup,” including the installation of a high-level containment lab, that “rang a lot of alarms,” Yassif says. Russia has insisted it is modernizing the labs to develop defenses against bioweapons. Arms control analysts also believe Russia is helping North Korea evade sanctions to acquire biotech equipment.

The ranks of potential violators is expanding, from nations to terrorist groups, as new technologies make it easier to manipulate life. “Emerging technologies might lower the barrier to entry,” says Bonnie Jenkins, who served as undersecretary of state for arms control and international security affairs in former President Joe Biden’s administration.

Potential acts of subterfuge also remain a major worry. One nightmare scenario Weber envisions is Russia unleashing a pathogen via the Port of Rotterdam in the Netherlands, Europe’s largest, then claiming it was an accidental lab release from nearby Erasmus University.

To reduce risk, NTI last year launched the International Biosecurity and Biosafety Initiative for Science to safeguard biotechnology from misuse. The initiative has begun to offer free screening software to DNA synthesis companies to help them avoid selling genes for, say, new bacterial toxins. The screening tool can pick up dangerous new proteins designed by AI, Microsoft’s Bruce Wittmann and colleagues reported in December 2024 on bioRxiv. But novel sequences generated by a pioneering technique that creates proteins from noncanonical amino acids—beyond the 20 protein building blocks found in nature—might slip past. Such genetic code expansion is “the risk I’m shouting about right now,” says Katarzyna Adamala, a synthetic biologist at the University of Minnesota Twin Cities.

In unpublished work, Adamala and colleagues sent the sequence of a harmless protein, a luciferase, to a collaborator at a DNA synthesis company and asked him to alter his screening database to flag it as a sequence of concern. Then, they sent a version of the sequence that coded for a luciferase built in part from noncanonical amino acids, which passed the screen. They repeated that with a couple other benign sequences, then with pathogenic sequences without ordering the DNA. All evaded the screen. “We were able to fly under the radar,” she says, suggesting “we need to change how screening works.”

BWC backers hope that in talks scheduled for coming months treaty parties will hammer out a mechanism for triggering inspections. “You don’t need inspections to be perfect. You just need them to raise the cost of cheating,” says Christopher Park, an arms control expert who last week retired from the U.S. Department of State. “Moving, hiding, or destroying material can destroy a bioagent,” adds Sonia Ben Ouaghram-Gormley, a bioweapons expert at George Mason University. “Even if a program resumes later, you’ve delayed it. In biology, that’s a big win.”

Any progress, however, will come only after treaty parties move past the ill will generated by last year’s meeting. And U.S. President Donald Trump’s new administration has not yet articulated a stance on the BWC. The general mood in nonproliferation circles, Jenkins says, is that “political will is still lacking” to make significant changes.

For now, the taboo that bioweapons represent to much of humanity may be the strongest safeguard against their use.

doi: 10.1126/science.zah2mbg

Richard Stone contributes to Science as its senior international correspondent with a focus on Asia. His writing has featured datelines from challenging reporting environments such as Cuba, Iran, and North Korea. Stone also serves as a special adviser, science diplomacy and engagement, for the Human Frontier Science Program organization. He has contributed to Discover, Smithsonian, and National Geographic magazines, and is the author of the nonfiction book Mammoth: The Resurrection of an Ice Age Giant. His overseas experience includes stints as a Fulbright Scholar at Rostov State University in Rostov-on-Don, Russia in 1995–96 and at Kazakh National University in Almaty, Kazakhstan in 2004–05.

SIGN UP FOR THE AWARD-WINNING SCIENCEADVISER NEWSLETTER

The latest news, commentary, and research, free to your inbox daily

Sign up

How Has Cosmology Changed From 2000 to 2025?
Ethan Siegel
Big Think - Starts With A Bang
25 years ago, our concordance picture of cosmology, also known as ΛCDM, came into focus. 25 years later, are we about to break that model?
April 8, 2025