When queried about the present state of biotechnology, Kevin Esvelt tends to sound more like a science skeptic than a pioneer of one of the most subversive genetic engineering technologies of all time.
"We are walking forwards blind," Esvelt told me recently, chastising his field. "We are opening boxes without thinking about consequences. We are going to fall off the tightrope and lose the trust of public."
If that happens, Esvelt said, lots of people who might have otherwise been saved are going to die.
Two years ago, Esvelt and his colleagues were the first to suggest that the gene-editing technology Crispr could be used to create what's known as a gene drive, a formidable tool that can be used to override natural gene selection during reproduction to ensure that a desired trait is passed down throughout generations. Using gene drives, scientists could potentially alter the entire population of a species. It is a proposition that is at once both spectacular and terrifying.
Ever since putting the idea out there, Esvelt has been hard at work warning the world just how dangerous it might be. He worries that a lab might let loose a genetically modified species that dramatically reshapes the natural world as we know it. Worse, he worries that there's nothing in the scientific community's current system of regulation that would prevent that from happening. Unless it's research happening on humans, gene drive work contained within the lab is not subject to very much regulatory scrutiny. The current system, he warns, is outmoded and "too risky."
"I occupy a strange position," said Esvelt, who at 33 heads up MIT's Sculpting Evolution Lab. "I am probably the foremost critic of genetic engineering and yet I am also someone at the forefront of the work I critique."
You are unlikely to find a work of modern science that more clearly demonstrates man's prowess over nature than the gene drive. In nature, the odds of either parent bestowing their own genetic characteristics upon their offspring are a 50-50 split. The gene drive is a genetic hack that subverts the very nature of reproduction, forcing propagating genes to prioritize the genetics of one parent over the other so that nearly every offspring receives a particular gene-driven trait. Scientists could, for example, engineer a mosquito to produce only male offspring, drastically limiting the ability of diseases like malaria or Zika to spread.
Gene drives exist in nature, and scientists have talked about replicating them for years, but Esvelt is the guy that figured out how to actually make it work. Using the highly precise Crispr-CAS 9 gene-editing system, Esvelt and his colleagues discovered they could make a desired trait cut and copy itself into every future generation. Esvelt's technique means that scientists now have the power to change a species' genetic make-up over just a few generations.
But the potential consequences are as perilous as the possibilities are glittering. As those malaria-fighting male mosquitos go on to mate with other mosquitos in the wild, they could also wind up wiping out the global mosquito population. What would happen if we accidentally killed off all the mosquitos? What else would die with them? What scientist alone is equipped to decide to take that kind of risk? This is the thought experiment with which Esvelt is concerned.
"It is impossible for a single scientist to anticipate the consequences of their work," he said. "We need a different approach."
The approach Esvelt is advocating for is to open the lab coat, so to speak.
When Esvelt and his colleagues published their 2014 paper floating the idea of using Crispr to make gene drives viable, it was just theoretical. They hadn't actually created a gene drive in the lab. But in the two years since their paper came out, gene drives have been demonstrated to work in four different species. That wasn't what Esvelt wanted. He had hoped other scientists would follow their lead, allowing fellow scientists and the public to vet experiment proposals and their potential consequences before putting the gene drive to use.
The root of Esvelt's concern is that scientists, like the rest of us, sometimes make mistakes. Those accidents can result in incredible discoveries, such as when Alexander Fleming discovered penicillin in 1928 by, essentially, leaving his lab window open. Other mistakes breed disaster. Many suspect, for example, that the 1977 flu epidemic was the result of a laboratory accident.
Esvelt would like to see scientists thinking not just about science, but about the ethics of science—about what might happen should their experiment go awry. This month Esvelt, along with MIT political scientist Kenneth Oye and Harvard social scientist Sam Weiss Evans, launched Responsive Science. The project is an online space where they hope competing scientists will publish research plans and proposals before going ahead with them. Then, he hopes, other experts in the field will weigh in, giving them feedback on their proposals.
Esvelt and colleagues published their first paper there last week, about a way to create gene drives with a kind of self-destruct button, which would, in theory, reduce the amount of harm a gene drive could do to a wild population.
"How much better would it be if all this research was just out there online?" he said. "Science would become more efficient and faster and we would have a better sense of what the consequences of the work would be."
The project, said Christopher Scott, a biomedical ethicist at Baylor University who is not associated with Esvelt, is just common sense.
"Scientists should be more upfront and engaged in their research in public," said Scott. But there are forces that work against that transparency. Because of the pressures on scientists to be the first to publish on a new discovery, scientists often keep their work secret until the last minute.
"Thoughtful and thorough engagement with issues like gene drives requires thinking well beyond the science involved, and that takes time, but that is time that traditional metrics of advancement in academia are against," Responsive Science founder Evans told me.
But Esvelt's concerns go beyond the potential catastrophic screw up: scientists, he believes, should not be in charge of making ethical decisions about technologies that could change the world. Imagine how the public would have responded had scientists of the Manhattan Project disclosed that they were working on a nuclear bomb. Would that research have moved forward? Would we have dropped bombs on Japan or entered the nuclear arms standoff that still exists today?
Esvelt wants to see scientists working more with local communities where their experiments might take place. As part of a plan to engineer mice that are immune to Lyme disease in an attempt to curb the disease's spread in the wild, Esvelt talked about the project at a community meeting in Nantucket, where he might eventually release such mice, before even going ahead with early stages of the experiment.
"To proceed with technology that will shape the global commons without public discussion is not only unethical, it's delusional," said Esvelt's collaborator, Oye.
Esvelt realized a more aggressive strategy for promoting caution among scientists was needed after he got wind of an unpublished paper by researchers who had created a gene drive to turn normal fruit flies yellow after a reporter forwarded a copy to his mentor, the eminent Harvard geneticist George Church.
Church's own lab, where Esvelt once worked, has long advocated openness when it comes to high-risk experiments. Church told me that it's important to "encourage multiple views of all proposals in advance." This he said, will "reduce corner-cutting," stimulate public engagement, and increase the chances of someone noticing something either "unexpectedly risky" or beneficial.
Both Esvelt and Church were horrified to find that the original paper—with the ominous-sounding title "The mutagenic chain reaction"—barely discussed the precautions necessary when working with gene drives. The researchers had been cautious in their own work to make sure that their mutant yellow flies didn't get out, but Esvelt felt that they hadn't taken enough steps to ensure containment. And most alarming, their paper offered little advice to researchers who might try to replicate their work about the importance of making sure the yellow flies never escape. (In the version of the paper published, the cautionary tone was significantly amped up.)
Evelt called the fruit fly experiment a "case study" in the dangers of closed-door science—when working at lab-level, it can be difficult to anticipate the consequences from a bird's-eye view.
"Fruit fly geneticists are not especially known for keeping their flies in the lab. It's nearly impossible," he said. “I thought to myself, 'If someone releases a gene drive into wild fruit flies, people will not trust us to do fruit fly genetics anymore.'"
Historically in genetic engineering, Mother Nature kills off mutant traits that escape into the wild. If a yellow fruit fly mated with a wild fruit fly, natural selection would likely weed out that yellow gene in favor of the color that has over time evolved to help the fly best survive its natural habitat. But add in a gene drive, and you could turn every fruit fly in the world a lovely shade of mustard.
For this reason, earlier this month, the federal National Academies of Sciences issued a report assessing the risks of gene drive technology. "The fast-moving nature of this field is both encouraging and a point of concern," the report said, asking that scientists take heed of "social, environmental, legal, and ethical considerations" to develop the technology responsibly. The report ultimately concluded that the potential benefits are too great to not proceed with “carefully controlled field trials."
Esvelt thinks that the report missed a major point: there's not really a great way to "control" experiments in the wild. At least not yet.
Efforts are already moving along to develop gene drives that could eliminate mosquitos carrying malaria, Zika, and dengue fever. The malaria project, being conducted by Imperial College London, has received more than $40 million in funding from the Gates Foundation and hopes to release its mosquitos in Africa by 2029.
The real hurdle to enacting Esvelt's open-source science plan will be convincing enough scientists to sign on. Esvelt hopes that journals and grant funders like the Gates Foundation will start requiring scientists to vet their experiments via a method like his before experimenting.
"It will take a while," said Church, whose lab is also committed to the project. "[But] all it takes is one fairly major player like the Gates Foundation to get on board in order to nudge it along."
(When asked about Esvelt's proposal, Scott Miller, deputy director of R&D for the Gates Foundation's malaria program, said that while they support "a transparent and inclusive approach to the development of guidance on gene drive research," it seems "premature to comment on specific proposals to regulate R&D.")
Other scientists, in fact, have previously argued for the opposite. At a meeting of the National Academies back in 2014 where Oye was presenting gene drives and public safety, he recalls one scientist who stood up and said, essentially, "don't tell the Muggles what's going on here with the dark arts."
And in a comment on a piece by Esvelt and Oye in Science arguing for openness, the neuroscientist David Gurwitz argued that all technical information concerning gene drives should be kept confidential. This, he said, was because "in the wrong hands" such technology might be used for nefarious purposes.
That risk, Esvelt believes, is far smaller than the risk that an unwitting scientist might accidentally make a devastating mistake—perhaps not a mistake large enough to wreck the ecosystem, but certainly one large enough to stymy scientific progress.
"We haven't seriously screwed up in the laboratory yet. Sometimes that surprises me," he said. "And we can't afford that right now. We need science. Our world is fundamentally unsustainable and we need to invent our way out of it. If we misuse our power, we lose the trust. That is the tightrope we walk."