Of the dozens of hormones found in the human body, oxytocin might just be the most overrated. Linked to the pleasures of romance, orgasms, philanthropy, and more, the chemical has been endlessly billed as the “hug hormone,” the “moral molecule,” even “the source of love and prosperity.” It has inspired popular books and TED Talks. Scientists and writers have insisted that spritzing it up human nostrils can instill compassion and generosity; online sellers have marketed snake-oil oxytocin concoctions as “Liquid Trust.”

But as my colleague Ed Yong and others have repeatedly written, most of what’s said about the hormone is, at best, hyperbole. Sniffing the chemical doesn’t reliably make people more collaborative or trusting; trials testing it as a treatment for children with autism spectrum disorder have delivered lackluster results. And although decades of great research have shown that the versatile molecule can at times spark warm fuzzies in all sorts of species—cooperation in meerkats, monogamy in prairie voles, parental care in marmosets and sheep—under other circumstances, oxytocin can turn creatures from rodents to humans aggressive, fearful, even prejudiced.

Now researchers are finding that oxytocin may be not only insufficient for forging strong bonds, but also unnecessary. A new genetic study hints that prairie voles—fluffy, fist-size rodents that have long been poster children for oxytocin’s snuggly effects—can permanently partner up without it. The revelation could shake the foundations of an entire neuroscience subfield, and prompt scientists to reconsider some of the oldest evidence that once seemed to show that oxytocin was the be-all and end-all for animal affection. Cuddles, it turns out, can probably happen without the classic cuddle hormone—even in the most classically cuddly creatures of all.

[Read: The weak science behind the wrongly named moral molecule]

Oxytocin isn’t necessarily obsolete. “This shouldn’t be taken as, ‘Oh, oxytocin doesn’t do anything,’” says Lindsay Sailer, a neuroscientist at Cornell University. But researchers have good reason to be a bit gobsmacked. For all the messy, inconsistent, even shady data that have been gathered from human studies of the hormone, the evidence from prairie voles has always been considered rock-solid. The little rodents, native to the midwestern United States, are famous for being one of the few mammal species that monogamously mate for life and co-parent their young. Over many decades and across geographies, researchers have documented how the rodents nuzzle each other in their nests and console each other when stressed, how they aggressively rebuff the advances of other voles that attempt to homewreck. And every time they checked, “there was oxytocin, sitting in the middle of the story, over and over again,” says Sue Carter, a behavioral neurobiologist who pioneered some of the first studies on prairie-vole bonds. The molecular pathways driving the behaviors seemed just as clear-cut: When triggered by a social behavior, such as snuggling or sex, a region of the brain called the hypothalamus pumped out oxytocin; the hormone then latched on to its receptor, sparking a slew of lovey-dovey effects.

Years of follow-up studies continued to bear that thinking out. When scientists gave prairie voles drugs that kept oxytocin from linking up with its receptor, the rodents started snubbing their partners after any tryst. Meanwhile, simply stimulating the oxytocin receptor was enough to coax voles into settling down with strangers that they’d never mated with. The connection between oxytocin and pair bonding was so strong, so repeatable, so unquestionable that it became dogma. Zoe Donaldson, a neuroscientist at the University of Colorado at Boulder who studies the hormone, recalls once receiving dismissive feedback on a grant because, in the words of the reviewer, “We already know everything that there is to know about prairie voles and oxytocin.”

So more than a decade ago, when Nirao Shah, a neurogeneticist and psychiatrist at Stanford, and his colleagues set out to cleave the oxytocin receptor from prairie voles using a genetic technique called CRISPR, they figured that their experiments would be a slam dunk. Part of the goal was, Shah told me, proof of principle: Researchers have yet to perfect genetic tools for voles the way they have in more common laboratory animals, such as mice. If the team’s manipulations worked, Shah reasoned, they’d beget a lineage of rodents that was immune to oxytocin’s influence, leaving them unfaithful to their mates and indifferent to their young—thereby proving that the CRISPR machinery had done its job.

That’s not what happened. The rodents continued to snuggle up with their families, as if nothing had changed. The find was baffling. At first, the team wondered if the experiment had simply failed. “I distinctly remember sitting there and just being like, Wait a sec; how is there not a difference?” Kristen Berendzen, a neurobiologist and psychiatrist at UC San Francisco who led the study, told me. But when three separate teams of researchers repeated the manipulations, the same thing happened again. It was as if they had successfully removed a car’s gas tank and still witnessed the engine roaring to life after an infusion of fuel. Something might have gone wrong in the experiments. That seems unlikely, though, says Larry Young, a neuroscientist at Emory University who wasn’t involved in the new study: Young’s team, he told me, has produced nearly identical results in his lab.

The explanations for how decades of oxytocin research could be upended are still being sussed out. Maybe oxytocin can attach to more than one hormone receptor—something that studies have hinted at over the years, Carter told me. But some researchers, Young among them, suspect a more radical possibility. Maybe, in the absence of its usual receptor, oxytocin no longer does anything at all—forcing the brain to blaze an alternative path toward affection. “I think other things pick up the slack,” Young told me.

That idea isn’t a total repudiation of the old research. Other prairie-vole experiments that used drugs to futz with oxytocin receptors were performed in adult animals who grew up with the hormone, says Devanand Manoli, a psychiatrist and neuroscientist at UCSF who helped lead the new study. Wired to respond to oxytocin all through development, those rodent brains couldn’t compensate for its sudden loss late in life. But the Stanford-UCSF team bred animals that lacked the oxytocin receptor from birth, which could have prompted some other molecule, capable of binding to another receptor, to step in. Maybe the car never needed gas to run: Stripped of its tank from the get-go, it went all electric instead.

It would be easy to view this study as yet another blow to the oxytocin propaganda machine. But the researchers I spoke with think the results are more revealing than that. “What this shows us is how important pair bonding is,” Carter told me—to prairie voles, but also potentially to us. For social mammals, partnering up isn’t just sentimental. It’s an essential piece of how we construct communities, survive past childhood, and ensure that future generations can do the same. “These are some of the most important relationships that any mammal can have,” says Bianca Jones Marlin, a neuroscientist at Columbia University. When oxytocin’s around, it’s probably providing the oomph behind that intimacy. And if it’s not? “Evolution is not going to have a single point of failure for something that’s absolutely critical,” Manoli told me. Knocking oxytocin off its pedestal may feel like a letdown. But it’s almost comforting to consider that the drive to bond is just that unbreakable.

For all the legwork that public-health experts have done over the past few years to quash comparisons between COVID-19 and the flu, there sure seems to be a lot of effort nowadays to equate the two. In an advisory meeting convened earlier today, the FDA signaled its intention to start doling out COVID vaccines just like flu shots: once a year in autumn, for just about everyone, ad infinitum. Whatever the brand, primary-series shots and boosters (which might no longer be called “boosters”) will guard against the same variants, making them interchangeable. Doses will no longer be counted numerically. “This will be a fundamental transition,” says Jason Schwartz, a vaccine policy expert at Yale—the biggest change to the COVID-vaccination regimen since it debuted.

Hints of the annual approach have been dropping, not so subtly, for years. Even in the spring of 2021, Pfizer’s CEO was floating the idea of yearly shots; Peter Marks, the director of the FDA’s Center for Biologics Evaluation and Research, teased it throughout 2022. This past September, Joe Biden officially endorsed it as “a new phase in our COVID-19 response,” and Ashish Jha, the White House’s COVID czar, memorably highlighted the convenience of combining a flu shot and a COVID shot into a single appointment: “I really believe this is why God gave us two arms.”

Still, in today’s meeting, FDA officials were pushier than ever in their advocacy for the flu-ification of COVID vaccines. “We think that simplification of the vaccination regimen would contribute to easier vaccine deployment, better communication, and improved vaccine coverage,” Jerry Weir, the FDA’s director of the division of viral products, said at the meeting. The timing is important: After renewing the U.S.’s pandemic-emergency declaration earlier this month, the Biden administration seems set to allow its expiration this coming April. That makes the present moment awfully convenient for repackaging a chaotic, crisis-caliber vaccination paradigm as a scheduled, seasonal, normal-seeming one. A once-a-year strategy, modeled on a routine recommendation, suggests that “we’re no longer in emergency mode,” says Maria Sundaram, a vaccine researcher at the Marshfield Clinic Research Institute. Or at least, that’s the message that the public is likely to hear.

But federal regulators may be trying to fit a COVID-shaped peg into a flu-shaped hole. The experts I spoke with largely agreed: Eventually, someday, annual autumn shots for COVID “will probably be sufficient,” says Gregory Poland, a vaccinologist at Mayo Clinic. “Are we ready for that yet? I’m not sure that’s the case at all.”

Even in the short term, COVID-vaccination tactics need a revamp. “It’s clear above all that the current approach isn’t working,” Schwartz told me. Despite abundant supply, demand for COVID boosters in the U.S. has been abysmal—and interest seems to be declining with each additional dose. Last fall’s bivalent shot has reached the arms of only 15 percent of Americans; even among adults over 65—a majority of whom sign up for flu shots each fall—the vaccination rate hasn’t yet reached 40 percent.

For most of the time that COVID shots have been around, figuring out when to get them has been a hassle, with different guidelines and requirements that depend on age, sex, risk factors, vaccination history, and more. Pharmacies have had to stock an absurd number of vials and syringes to accommodate the various combinations of brands and dose sizes; record-keeping on flimsy paper cards has been a total joke. “I do this for a living, and I can barely keep track,” Schwartz said. Recommendations on the proper timing and number of doses have also changed so many times that many Americans have simply checked out. After the bivalent recipe debuted, polls found that an alarming proportion of people didn’t even know the shot was available to them.

Streamlining COVID-vaccine recommendations will remove a lot of that headache, Sundaram told me. Most people would need to keep only one mantra in mind—one dose, each fall—and could top off their flu and COVID immunity at the same time. Burdens on pharmacies and clinics would be lower, and communication would be far easier—a change that could make an especially big difference for those with children, among whom COVID-vaccine uptake has been the lowest. “It’ll be more scheduled, more systematic,” says Charlotte Hobbs, a pediatric infectious-disease specialist at the University of Mississippi Medical Center. COVID shots could simply be offered at annual well-child visits, she told me. “It’s something we already know works well.”

The advantages of a flu-ified COVID shot aren’t just about convenience. If we have to shoehorn COVID vaccines into an existing paradigm, Sundaram told me, influenza’s is the best candidate. SARS-CoV-2, like the flu, is excellent at altering itself to dodge our defenses; it spreads readily in winter; and our immunity to infection tends to fade rather quickly. All of that adds up to a need for regularly updated shots. Such a system has been in place for decades for the flu: At the end of each winter, a panel of experts convenes to select the strains that should be targeted by the next formulation; manufacturers spend the next several months whipping up big batches in time for an autumn-ish rollout. The pipeline depends on a global surveillance system for flu viruses, as well as regular surveys of antibody levels in the community to suss out which strains people are still protected against. The premise has been so well vetted by now that researchers can skip the chore of running large-scale clinical trials to determine the efficacy and safety of each new, updated recipe.

But a seasonal strategy works best for a seasonal virus—and SARS-CoV-2 just isn’t there yet, says Hana El Sahly, an infectious-disease physician at Baylor College of Medicine. Though flu viruses tend to hop between the globe’s hemispheres, alternately troubling the north and the south during their respective cold months, this new coronavirus has yet to confine its spread to one part of the calendar. (Marks, of the FDA, tried to address this concern at today’s meeting, asserting that “we’re starting to see some seasonality” and that fall was indeed the very sensible for an annual rollout.) SARS-CoV-2 has also been spitting out concerning variants and subvariants at a faster rate than the flu (and flu shots already have a hard time keeping up with evolution). The FDA’s new proposal suggests picking SARS-CoV-2 variants in June to have a vaccine ready by September, a shorter timeline than is used for flu. That still might not be fast enough: “By the time we detect a variant, it will have ripped through the global population and, in a few more weeks, died down,” El Sahly told me. The world got a preview of this problem with last year’s bivalent shot, which overlapped with the dominance of its target subvariants for only a couple of months. A flu model for COVID would make more sense “if we had stable, predictable dynamics,” says Avnika Amin, a vaccine epidemiologist at Emory University. “I don’t think we’re at that point.”

Murkiness around vaccine effectiveness makes this transition complicated too. Experts told me that it’s gotten much more difficult to tell how well our COVID vaccines are working, and for how long, fueling debates over how often they should be given and how often their composition should change. Many people have now been infected by the virus multiple times, which can muddy calculations of vaccine effectiveness; better treatments also alter risk profiles. And many researchers told me they’re concerned that the data shortcuts we use for flu—measures of antibodies as a proxy for immune protection—just won’t fly for COVID shots. “We need better clinical data,” El Sahly told me. In their absence, the hasty adoption of a flu framework could lead to our updating and distributing COVID shots too often, or not often enough.

A flu-ish approach also wouldn’t fix all of the COVID vaccines’ problems. Today’s discussion suggested that, even if a new COVID-shot strategy change goes through, officials will still need to recommend several different dose sizes for several different age groups—a more complex regimen than flu’s—and may advise additional injections for those at highest risk. At the same time, COVID shots would continue to be more of a target for misinformation campaigns than many other vaccines and, at least in the case of mRNA-based injections, more likely to cause annoying side effects. These issues and others have driven down interest—and simply pivoting to the flu paradigm “is not going to solve the uptake problem,” says Angela Shen, a vaccine-policy expert at Children’s Hospital of Philadelphia.

Perhaps the greatest risk of making COVID vaccines more like flu shots is that it could lead to more complacency. In making the influenza paradigm a model, we also threaten to make it a ceiling. Although flu shots are an essential, lifesaving public-health tool, they are by no means the best-performing vaccines in our roster. Their timeline is slow and inefficient; as a result, the formulations don’t always match circulating strains. Already, with COVID, the world has struggled to chase variants with vaccines that simply cannot keep up. If we move too quickly to the fine-but-flawed framework for flu, experts told me, it could disincentivize research into more durable, more variant-proof, less side-effect-causing COVID shots. Uptake of flu vaccines has never been stellar, either: Just half of Americans sign up for the shots each year—and despite years of valiant efforts, “we still haven’t figured out how to consistently improve that,” Amin told me.

Whenever the COVID-emergency declaration expires, vaccination will almost certainly have to change. Access to shots may be imperiled for tens of millions of uninsured Americans; local public-health departments may end up with even fewer resources for vaccine outreach. A flu model might offer some improvements over the status quo. But if the downsides outweigh the pluses, Poland told me, that could add to the erosion of public trust. Either way, it might warp attitudes toward this coronavirus in ways that can’t be reversed. At multiple points during today’s meeting, FDA officials emphasized that COVID is not the flu. They're right: COVID is not the flu and never will be. But vaccines can sometimes become a lens through which we view the dangers they fight. By equating our frontline responses to these viruses, the U.S. risks sending the wrong message—that they carry equal threat.