As Earth enters the Anthropocene epoch, its biodiversity wobbles on the precipice of disaster—and island species have been hit especially hard. About 80 percent of recorded extinctions have occurred on islands and 40 percent of the world's endangered and threatened species are island dwellers. Researchers say the leading cause of these extinctions is invasive rodents—rats and mice that stowed away on ships, then quickly populated islands where they have no natural predators and often find a buffet of things like eggs and baby wildlife. Whereas there are several ways to clear such invaders, the most effective has been rodenticides. But these poisons can neither be deployed effectively on islands with large human populations nor where residents disapprove of their use. And poisons do not discriminate, killing along with unwanted pests the native species they are meant to protect.

But now a controversial new strategy called gene drive offers a brutally efficient solution by introducing genetically modified organisms designed to spread a chosen trait—such as producing infertile offspring—throughout a wild population. Scientists, government officials and other interested parties debated the idea last week at the International Union for Conservation of Nature's (IUCN) World Conservation Congress in Honolulu.

“A gene drive works by distorting inheritance in its favor,” says Congress participant and biochemist Kevin Esvelt of Massachusetts Institute of Technology, one of the field's leading researchers. When two organisms reproduce, their offspring naturally have a 50 percent chance of inheriting any particular gene from either parent. A gene drive increases those chances to more than 50 percent. Over successive generations that gene “drives through” a population until most individuals possess it. Gene drives already occur naturally in nearly every species we know of, Esvelt says, and researchers have been dreaming up ways to take advantage of the phenomenon for nearly a century. But now scientists finally have a tool that allows them to harness the power of gene drives more precisely and efficiently than ever: CRISPR–Cas9, the affordable and easy-to-use gene-editing tool that allows scientists to modify an organism's genome with extreme precision.

In 2014 Esvelt and others published a paper outlining for the first time the potential ways in which CRISPR could be used for gene drives. “It makes it feasible in pretty much any sexually reproducing organism,” Esvelt says. For example, a researcher could modify a rat genome to contain a broken version of a gene necessary for female fertility, and introduce some of these rats to an island. They would mate with wild rats, and all of their offspring would inherit that broken gene. When those offspring then mated with more wild rats, their offspring would all inherit it as well. Given enough time, the entire female population would become infertile and the rats would die out. “Removing invasive species is almost like hitting an island's reset button,” says Heath Packard, a spokesperson for Island Conservation. The nongovernment organization’s stated mission is to prevent extinctions by removing invasive species from islands.

But this is easier said than done. “The size and scale of islands we're attempting is pushing the boundaries of the current technology,” which typically involves using anticoagulant rodenticides, says Karl Campbell, Island Conservation’s project director. Rodents have been introduced to some 80 percent of the world's islands, and Campbell says that current practices are only useful for around 10 percent of them. In researching and assessing potential new strategies, Campbell and his colleagues have made an argument for the use of gene drives on islands. Drive systems could be implemented for female infertility or alternatively to make mice more likely to produce infertile male offspring, and the resulting all-male rodent populations would eventually die out. Such systems could conceivably be used on islands as large as those of New Zealand, which recently announced its intention to rid itself of invasive predators by 2050.

Meanwhile, Floyd Reed, a biologist at the University of Hawai'i at Mnoa, has been working on a different kind of drive system called underdominance to prevent the Culex mosquitoes introduced to Hawaii from spreading avian malaria to endangered birds, including the Hawaiian honeycreeper. This technique involves releasing enough genetically modified mosquitoes into the environment until they comprise more than 50 percent of the overall population. Once the population reaches that threshold, natural selection works to favor the modified mosquitoes. Eventually the wild mosquitoes die out, leaving the modified ones in their place. In such a system the mosquitoes could be engineered to be incapable either of transmitting the malaria parasite or of reproducing. So far Reed has established a proof of concept in Drosophila fruit flies, a commonly used model species in laboratories across the world. But establishing an underdominance-based gene drive in Culex mosquitoes in the lab has been slow going. Progress might not even be fast enough to save the honeycreeper. Unless something can be done, he says, “in five years, we'll lose at least another species.”

A gene drive solution obviously comes with some serious concerns. Fortunately, an underdominance-based system should make it fairly easy to return future generations to their original, unmodified state by releasing enough wild mosquitoes back into the population. But some of the most powerful forms of gene drives are hard to control or reverse, and without the proper biosecurity mechanisms they could theoretically spread beyond the target population to impact an entire species. “No one should even be building a drive system like this to solve a conservation problem. It's just too early. We don't know enough,” Esvelt says, adding, “I'm probably the foremost scientific critic of gene drives even though I'm a leader in the field.”

Reed, too, sounds a careful note. “There is potential, but there are some technologies that we need to be very cautious about,” he says. A group of high-profile international activists including Jane Goodall and David Suzuki released a letter, coinciding with the IUCN meeting, openly opposing the release of what they call “genocidal genes” on ecological and moral grounds. Signatories include Claire Cummings, an author and former U.S. Department of Agriculture lawyer. “The existing regulatory framework cannot in any way intervene in this technology,” she says, adding, “there are some strong moral and ethical issues here, but this is very new. This is a real opportunity for us to ask the right questions.”

Indeed, when Esvelt wrote his 2014 paper outlining possible uses for the technology, his team also published a second paper explicitly calling for regulatory reform both in the U.S. and abroad. Arizona State University biologist Jim Collins, who convened a group of experts to explore the potential applications for gene drive technology earlier this year, echoed Esvelt, saying, "we need different arrangements. The [regulations] that are in place right now are not sufficient."

Luckily there is still time. Even if field testing gene drives safely was possible, it is at least five to 10 years away. “This is a technology that's going to be tremendously powerful, and any tremendously powerful technology needs to be handled very carefully and developed in the open light of day,” Esvelt says. “But it’s not here yet.”