I was planning to spend Sunday doing gain of function research with some friends, but maybe we’ll go for a walk instead.
A lot of meth lab operators have switched to gain of function in the public interest.
Fauci (b. 1940) has supported gain of function research for two decades. In December 2011 he co-wrote a Washington Post op-ed with Francis Collins (b. 1950) and Gary Nabel titled “A flu virus risk worth taking,” defending the position that generating a potentially dangerous virus in the laboratory can yield important information, while conceding the risk-benefit question has no simple answer. At a press conference on the Fouchier H5N1 work he acknowledged a researcher might infect himself and seed an outbreak, then said the benefits far outweigh the risk. He also welcomed the lifting of the funding pause at a 2018 NIAID conference. His denial before Rand Paul (b. 1963) that NIH funded gain-of-function work in Wuhan rests on the narrow regulatory definition, under which the category covers a pathogen already determined to be a potential pandemic pathogen, and the bat coronaviruses in the EcoHealth grant had received no such determination. So the two decades are real, and the denial turns on a definition rather than on the underlying conduct.
Virologists lean hard toward zoonosis for Covid origins. The 2024 Global Catastrophic Risk Institute survey put the question to 168 scientists across 47 countries. Respondents assigned an average 77 percent probability to natural spillover, 21 percent to a research accident, and 2 percent to something else. A quarter put zoonosis at 96 percent or higher. One in five gave a probability of 50 percent or more to something other than natural spillover. The survey found no statistically significant difference between virologists and epidemiologists.
Compare the public. Morning Consult in February 2023 found 44 percent of American adults picked a Wuhan lab, a share stable since mid-2021. Quinnipiac had it 64 to 22 for a lab that March; an Economist/YouGov poll ran 66 to 16.
Under both origin scenarios, a majority of responding experts favored more aggressive governance. Nine in ten said the next pandemic will most likely come from natural spillover, and only 8 percent picked a research accident. Believing nature did it last time turns out to be compatible with wanting tighter rules on the labs.
Simon Wain-Hobson of the Pasteur Institute writes a weekly column against gain of function research. Richard Ebright of Rutgers testified to the Senate that American oversight lapses may have caused the pandemic and that the Wuhan work met the criteria for the pause and for federal risk-benefit review and got neither. Marc Lipsitch at Harvard began questioning the risk-benefit balance in 2012. Jesse Bloom, an evolutionary virologist who accepts the zoonotic lean, still urges continued investigation with an open mind alongside steps to reduce the risk of both research-related and natural outbreaks. This is a fight inside the field.
That said, proximity predicts position. A PNAS study of 378 scientists who signed petitions for or against found that supporters tended to specialize in directly related fields such as virology and immunology, and that a scientist’s probability of supporting the work dropped when his collaborators signed the opposing petition. Support clusters where the work gets done.
Much of the apparent expert-public gap dissolves on definitions. Virologists apply the term to a broad range of experiments, most of them low risk, which is why some prefer “gain-of-function research of concern” or “enhanced potential pandemic pathogen” research for the dangerous subset. Ask a virologist whether he supports gain-of-function research and he hears a question about ordinary molecular biology. Ask a voter and he hears a question about building pandemic viruses. They answer different questions and then disagree about the answers.
As Science put it, if a lab accident killed millions it would stain virology and invite new restrictions on the field. That does not settle whether the majority view is correct. It does mean the people best equipped to judge the evidence also carry the largest stake in the verdict, which is the sort of thing you would want disclosed in any other domain.
The intelligence community splits. Four elements and the National Intelligence Council favored natural exposure with low confidence, Energy and the FBI favored a laboratory incident, and two agencies including CIA reached no conclusion. On January 25, 2025, CIA shifted to favoring a research-associated origin, a low-confidence judgment reached by reviewing existing material rather than any new intelligence. On policy the critics have won. Trump signed an executive order on May 5, 2025 halting federal support for what the administration calls dangerous gain-of-function research at home and abroad, and HHS announced a government-wide policy in late July 2026 barring federal funding for it. The New Atlantis account of how that happened is the one I would read on this: M. Anthony Mills argues the field treated the governance of this research as a technical problem for specialists when it was always a political question requiring public accountability, and so brought the crisis of public doubt on itself.
