On an early spring morning, a humming drone hovered over a small town in Bahia, Brazil. Three hundred feet above ground, a small canister clicked open, ejecting its contents into the mouth of the release mechanism below. For a moment, there was silence. Then, a swarm of mosquitoes, freshly awoken from icy slumber, stretched their wings and took flight.

Each specimen was male, single and ready to mingle—and if all went as planned, the buzzing horde of eager virgins would steadily infiltrate the local mosquito population, coupling up with thousands of lucky ladies in the days to come.

Considering that there are about 100 species of mosquito that carrying deadly human pathogens—including parasites that cause malaria, as well as Zika, dengue and West Nile viruses—this may sound like the horrifying start to an apocalyptic science fiction film à la Outbreak . But it’s quite the opposite: The mosquitoes unleashed in this experiment may be some of the best weapons against the spread of infectious disease.

For the past two years, a team of scientists and engineers from WeRobotics and the International Atomic Energy Agency (IAEA) has been testing new ways to disperse sterile male mosquitoes into regions where these deadly diseases run rampant. The researchers’ first drone-based trial run , conducted this past spring in Brazil, yielded promising results, and they’re already gearing up for more.

“This is a really exciting step forward,” says Kelsey Adams , a mosquito biologist at the Harvard School of Public Health who is not affiliated with the work. “With innovative techniques such as these, we can expand the areas into which we’re releasing [modified mosquitoes].”

The end goal is simple: Crowd out fertile males with eunuchs, and watch the numbers of potential disease-carrying mosquitoes plunge. This so-called sterile insect technique has already yielded success in agricultural pests like fruit flies, and in other bugs that ferry illness from person to person such as tsetse flies. Use of the technology in mosquitoes, however, is a somewhat newer phenomenon.

When it comes down to it, the sterile insect technique is a game of numbers. Wild populations must be inundated with lab-grown duds, sometimes in ratios upwards of 10 or more sterile males for every fertile local. And one-off dumping won’t do the trick: A region needs to be flooded again and again , until populations of native mosquitoes are driven down to negligible levels (and even then, it can remain an odious exercise in maintenance). What’s more, some species of mosquito, including the Aedes aegypti mosquitoes that transmit Zika, dengue and yellow fever, are classic couch potatoes, often traveling no more than a couple hundred feet in a lifetime. This further ups the difficulty of ensuring widespread coverage.

Mosquitoes can now be bred and sterilized en masse in a range of laboratory environments—a quick zap of radiation is enough to severely damage insect sperm. However, the process of safely packaging, transporting and delivering these sterile soldiers to the sites where they will do their dirty work is its own hurdle. Most efforts so far have involved human-powered ground releases from vehicles—but bumpy truck rides on unfinished roads inevitably jostle their precious cargo, and many regions riddled with disease are inaccessible by car.

Instead, researchers now look skyward.

“Drones are really a game changer,” says Jürg Germann , WeRobotics’ lead engineer.

The technology is surprisingly ubiquitous: For a few thousand dollars, drones can be purchased and transformed into mosquito chauffeurs. Compared to clunky cars, drones are at least five to 10 times more efficient at dispersing mosquitoes, says Germann. What’s more, drones are aerodynamic, reusable and completely unencumbered by roads (or lack thereof). Wherever there is sky, a drone can go, with hibernating mosquitoes in tow.

Previous work with fruit flies has used high-altitude aircrafts. But unlike planes, drones can fly low to the ground, ensuring more precision and control—and minimizing the damage fragile bugs might sustain as they’re vaulted off the craft. Best of all, drones have no need for error-prone pilots: The researchers can set their course at home base and wave goodbye.

After a year of prototyping, WeRobotics and IAEA took their efforts into the field. They set their sights first on Brazil, where a Zika epidemic, spread by Aedes aegypti mosquitoes, led to thousands of birth defects in 2015 and 2016.

Over the course of three trials, the researchers dispersed a total of 284,200 sterile male Aedes aegypti mosquitoes around the Brazilian community of Carnaíba do Sertão in March of this year. With the drones flying at full speed, the researchers were able to cover thousands of square feet in a matter of minutes—and over 90 percent of the airdropped mosquitoes appeared to stick the landing.

Surviving the perilous drop was just the beginning—but IAEA mosquito expert Jérémy Bouyer was pleased to see that these drone-derived dudes held their own against their fertile counterparts, fathering about one sterile egg for every viable egg produced by a wild male. Bouyer is optimistic that with more finagling, the numbers will continue to climb .

At such an early stage, it’s challenging to assess the long-term impact of these infertile insect blitzes. But epidemics hit when you least expect them—and insect control is all about nipping disease in the bud. WeRobotics and IAEA are already planning more trials in the months and years to come. In the meantime, Germann and his team are working on increasing each drone’s capacity and minimizing mosquito mortality. Eventually, the researchers hope to pass a better version of their technology onto local experts through an intensive training program, creating a sustainable and self-sufficient system of dispersal. By 2020, WeRobotics plans to have about 30 drone-dispatching stations worldwide.

“We’re not just out to throw technology at the problem—that’s not impact,” says WeRobotics co-founder Patrick Meier . “The drones should be the heroes in this story. Not the Western organizations.”

As the technology continues to progress, still more doors may open. In the fight against disease-carrying insects, drones aren’t good for just propagating packages of pests. Unmanned aircraft have already been used to map mosquito breeding sites —which are cumbersome to spot and track with the naked eye—allowing researchers to more easily study mosquito behavior and dispersal . Additionally, from their lofty vantage point, drones are an excellent tool for spraying insecticides .

Even within the scope of mosquito-toting drones, there is additional room for growth, says Adams, who studies the reproductive behavior of Anopheles mosquitoes, which can carry parasites that cause malaria , under the supervision of infectious disease researcher Flaminia Catteruccia . The drones are certainly not species-specific, and Bouyer, Germann and Meier are optimistic that Anopheles and other mosquitoes could be viable candidates for dissemination in the future.

In fact, when it comes to Anopheles mosquitoes, Adams says, drones could be even more of a boon. Anopheles aren’t lethargic lumps like their Aedes cousins. Airdropping these malaria mongers could be more bang for your buck, because fewer mosquitoes can cover a larger geographic area.

What’s more, there’s evidence that most Anopheles (and some Aedes ) females are largely monogamous; in fact, for the ladies of some Anopheles species, mate choice is a literal once-in-a-lifetime decision. And hormonal and behavioral research in Catteruccia’s group and others has shown that, even in a swarm of mating mosquitoes, a handful of hunks are disproportionately successful. This means a couple things: First, most male mosquitoes will, sadly, die virgins. Second, and perhaps more importantly, females are likely cueing into some indicator of male machismo as they make their way through the crowd. Someday, Adams says, scientists may be able to goad female mosquitoes into preferring sterile males to fertile ones, given the right incentives.

But sterile insect technique is only one of many strategies through which to target the reproductive cycle of these bloodthirsty bugs. And though it’s been highly effective in many contexts, this strategy is not without its drawbacks.

“One of the biggest problems is that it’s not self-perpetuating,” Adams explains. “You often end up needing more mosquitoes than you’d think.”

For one thing, setting up this laboratory breeding and sterilization centers worldwide would be a formidable feat. During their small, single-community trial in Brazil, WeRobotics and IAEA reared over 700,000 mosquitoes —a number that would have to be scaled up immensely to meet even a fraction of global need. Additionally, isolating a male-only population for sterilization and release isn’t as easy as it sounds, and mistakes can be extremely costly. Females are the bloodsuckers of the bunch—and thus the envoys of disease. The accidental release of a population that is even 1 percent female could actually worsen an epidemic, says Adams.

Luckily, many alternatives exist. One option involves manufacturing mosquitoes that can pass lethal genes onto their offspring—another form of sneaky birth control. This technology can be especially powerful when it’s engineered alongside a “ gene drive ”—essentially, a genetic element that strongarms its way into all offspring, no matter which parent carries the trait. Gene drives thus spread through populations at an accelerated rate, making them more efficient than strategies like the sterile insect technique: A smaller number of insects can dominate a wild population, somewhat relieving the need for mass mosquito production.

One concern that some scientists have raised in recent years is that several of these methods aim to completely wipe out certain species of mosquitoes. And disease-related or not, an extinction is an extinction.

But around 3,500 species of mosquito roam the skies—and many of them have overlapping ecological functions, says Adams. “Eliminating one mosquito species won’t necessarily have huge environmental consequences, considering that there are thousands of them,” she explains. “But of course, we should still proceed with caution.”

One alternative to out-and-out genocide is to introduce immunity to parasites or viruses into a mosquito population. Coupled with gene drives, this technology could potentially create a lasting lineage of insects that are free to sup on blood to their hearts’ content—and blissfully free of disease.

Of course, these methods aren’t foolproof either. Just as bacteria, viruses and parasites develop resistance to drugs, mosquitoes can mutate their way out of gene drives and other types of DNA manipulations. Even if it happens at low rates, one individual could quickly propagate its genetic hiccups on to future generations, undoing years of effort.

Bouyer points out that mishaps with mutations can be circumvented by relying on the original irradiation-based sterile insect technique: It’s not easy to find a genetic workaround for a sexual partner’s infertility. Additionally, while sterile insect techniques have been used for decades, he adds, sophisticated and invasive genetic modifications might encounter more obstacles on the road to commercialization, given the stringent GMO regulations that exist in many countries.

No single mosquito control strategy is likely to be a panacea on its own. However, Bouyer says, in the future, some of these techniques could be used effectively in combination . As is the case with medicines and other drugs, it’s far more difficult for a population to develop resistance when it’s battling several opponents at once.

In any case, Meier is enthusiastic about the possibility of new passengers for Air Mosquito. “We’re just the limo,” says with a laugh. “Whatever mosquitoes go into the limo is up to other experts. On our end, as long as there are mosquitoes, the [drop] will work—regardless of how they’ve been modified.”