Animal testing, explained

It’s nearly impossible to go a day without benefitting from the suffering of animals. The ingredients in your toiletries and makeup; your medicine, vaccines, and implants; your cleaning supplies; the chemicals that helped grow your food — most of it was, at some point, tested on animals.
For centuries, the biological sciences have relied on animal testing. To figure out how a machine works, you need to disassemble it and check out its component parts. Understanding the living body, one of nature’s most complex and beautiful machines, is no different. Taking apart and fiddling with a toaster doesn’t hurt anyone, but dismantling a biological system certainly does.
Many scientists believe that experimenting on living animals is a necessary means of solving problems that affect both humans and animals. But these experiments often involve animals experiencing distress, whether from the side effects of an experimental drug, an intentionally inflicted illness, or simply their confined living situation. Some lucky lab animals get to spend their retirement in sanctuaries once they’re no longer needed. Most of the time, the animal dies, either as a direct consequence of the experiment or from euthanasia.
More often than not, animal research happens behind closed, locked, unmarked doors. That lack of transparency makes it difficult to know what to think about animal testing, and public opinion is tellingly divided. A 2018 Pew Research Center survey found that 47 percent of people in the US support the use of animals in scientific research, and 52 percent oppose it. Unlike climate change or reproductive health, where the parties are highly polarized, animal testing is one of few science-related policy issues where the attitudes of Republicans and Democrats are pretty similar: Both parties are split roughly 50-50.
Experimenting on animals places two seemingly good things — medical innovation and animal welfare — at odds. Even those who support animal research generally hold nuanced, conflicted beliefs about it, and questions about the nature and extent of animal testing are still hotly debated.
Brands frequently mislead consumers about animal testing involving their products with vague labeling, and alternative research methods aren’t as broadly applicable as some activist organizations imply. Meanwhile, research facilities often ban employees from sharing photos of lab animals without institutional approval and rarely let the media observe experiments for themselves.
After spending six years as a neuroscience PhD student working in a lab with monkeys, I left academia with the impression that animal testing is neither as well-managed or justified as regulators claim, nor as malicious as others fear. Government agencies are starting to direct funding toward finding alternatives to animal testing, but the use of animals is deeply embedded in biological sciences.
A world without lab animals may be possible, but we don’t live in it yet. Here’s what’s actually going on.
Before humans invented microscopes, universities, or even paper, we were using animals for medical research. Over two millennia ago, ancient Greek philosopher Aristotle dissected dozens of animal species to better understand their anatomy and argued that studying their bodies could teach us a lot about our own biology. Over four centuries later, Galen of Pergamon, one of the most pivotal characters in Western medical history, performed public surgeries on animals (especially monkeys) for science, providing a spectacle that attracted curious audiences.
Today, animal experimentation is widespread and conducted far from the public eye. It falls under two broad, semi-overlapping umbrellas: biomedical research (which aims to understand, prevent, and treat diseases, as well as uncover fundamental information about how bodies work) and toxicology, or testing the effects of chemicals (including everything from toothpaste and makeup to pesticides) on living things.
Humans generally don’t want to be proverbial guinea pigs for new medicines or consumer products. We’d rather know that things are safe before we put them anywhere near our bodies. Companies, whether they deal in cosmetics or pharmaceuticals, also don’t want to be liable for poisoning their customers.
People can participate in experiments that might harm them, but historically, at best, such projects have been difficult to administer. At worst, they have involved illegal human experimentation that cast a long, dark shadow over the field of medical research.
The Tuskegee syphilis study, for example, put hundreds of poor Black men with untreated syphilis through decades of invasive tests in exchange for hot meals and basic medical treatment, just to see how the disease would progress if left untreated. Effective treatments became available during the study, but researchers withheld them. Once the experiment’s scandalous history was publicly disclosed in 1972, the US government formalized basic ethical guidelines for human research and required Institutional Review Boards (IRBs) to approve studies on humans.
Today, many questions — like What is the lethal dose of this new drug? and Does this new surgical technique actually work? — can’t ethically be asked regarding humans without first being tested on a nonhuman subject.
For a long time, animals were the only alternative to humans available. To figure out the lethal dose of a new drug, scientists can give increasingly large amounts of it to mice and see what it takes to kill them. To test whether a brain implant actually relieves Parkinson’s symptoms, scientists do brain surgery on monkeys. Without computational models or cell cultures sophisticated enough to mimic the complicated interactions between organs, the options have historically been to use animals as a proxy or to drop or scale back your planned research.
We can only guess how many animals are being used in scientific experiments worldwide. The United States Department of Agriculture (USDA) publishes official reports on animal research every year, but they only include animals protected by the Animal Welfare Act (AWA), the federal law setting basic standards for the treatment and housing of certain farm animals and lab animals. The law covers dogs, cats, monkeys, guinea pigs, hamsters, pigs, rabbits, and sheep. In 2019, about 800,000 animals protected by AWA were used in research — 930,000, if you add those that lived in labs but were never included in a study.
Notably, the AWA doesn’t apply to mice and rats, which several studies estimate account for somewhere between 93 and 99 percent of all lab animals in the US. The AWA also excludes invertebrates like flies, worms, fish, and cephalopods like octopuses, whose intelligence makes them intriguing neuroscience subjects. The EU, which counts all vertebrates used in experiments, tallied about 10.6 million animals used in 2017. It’s harder to pin down a number in the US. Depending on who you ask, there might be 10 million rodents subjected to scientific experiments annually, or there might be 111 million. (Either way, it’s more than three times the number of rats in New York City.)
Rodents make appealing animal models for many scientists because they’re smart enough to learn simple tasks but are still socially regarded as pests; those who kill rats for a living don’t face the same kind of backlash as someone who, say, boasts about shooting a puppy. Nearly all mouse genes share functions with human genes, so at a basic level, their biology resembles ours. Mice only live for a year or two, enabling scientists to study things like chronic disease progression without waiting an entire human lifespan. And scientists can genetically alter mice in countless ways, knocking out or adding DNA to express diseases or make certain cell types glow under a microscope.
In some cases, a research question requires invasively studying a full, living biological system, but the gap between mice and humans is too wide. The USDA reported that 68,257 monkeys were used in 2019 to study subjects like SARS-CoV-2, Parkinson’s disease, and HIV, where physiological and cognitive similarity to humans was a priority. Those primates were mostly macaques and marmosets; the use of chimpanzees (our closest ape relative) is now banned in many countries, including the US.
But monkey research may not be viable much longer. While hundreds of monkey experiments are being funded by the NIH, there aren’t enough long-tailed macaques to go around. In a desperate attempt to keep up with skyrocketing demand, thousands of wild-caught monkeys are illegally imported to US research institutions from countries like Cambodia. Two years ago, the long-tailed macaque was listed as endangered for the first time. PETA petitioned the US government to protect the species under the Endangered Species Act, which could end their use in research altogether, but the request has yet to be approved.Most people are uncomfortable with the idea of experimenting on an animal so similar to us, including some of the scientists who do it. However, many scientists and policymakers agree that we still don’t have non-animal alternatives that can answer tough research questions involving interactions between organs. Researchers worry that the looming primate shortage in the US — engendered by transportation restrictions and therapeutic testing requirements and exacerbated by pandemic-era demands — will limit our ability to respond to public health emergencies.
Monkeys are traditionally recognized as the only nonhuman animals that react to drugs with human-specific targets, meaning that in some cases, their body’s reactions could uniquely predict whether a drug will be safe and effective for humans. During the first years of the Covid-19 pandemic, monkeys were considered so crucial to SARS-CoV-2 research that when the rhesus macaque supply dried up, scientists didn’t turn to cell cultures or computer models — they just looked for different monkeys.
You might not agree that this research justifies the nonconsensual use of highly intelligent animals; many don’t, for both ethical and scientific reasons. But it’s happening, and if you’ve been vaccinated or take medications, you’ve likely benefited from it.
The regulatory framework surrounding animal research is a tangled web of acronyms, committees, and working groups. Since the Animal Welfare Act was passed in 1966, the USDA has been in charge of enforcing it through inspections and annual reports.
In theory, researchers have to justify the use of animals in their work. To conduct an animal experiment, scientists in the US go through a review process with their Institutional Animal Care and Use Committee (IACUC), which decides whether animals are “necessary” and whether steps are being taken to minimize their pain.
IACUCs are mostly comprised of researchers who experiment on animals and the veterinarians who help them, strongly biasing committees toward approving animal experiments. In the US and elsewhere, scientists are subtly incentivized to use animals, even when they aren’t actually necessary. Academic journals tend to preferentially publish work with animal methods, and academic careers hinge on accumulating publications. These norms seep into the labs where animal experiments are performed. New animal researchers often receive explicit instructions on how to steer clear of animal rights activists, according to several researchers I spoke with while working as a neuroscientist (as well as my own experience).
This can make holding institutions accountable for animal welfare violations challenging. While researchers are required to report information about animals in their facilities, like what medical procedures they’ve received and when they’ve been fed, they are told to keep these reports “minimal, but complete.” In other words: Avoid including photos, videos, or graphic descriptions that could enrage activists or entice the media.
There also isn’t a clear legal definition for “animal cruelty” in research settings beyond violations of the basic standards outlined by the Animal Welfare Act. This leaves some room for interpretation about what is acceptable and what would constitute illegal treatment. The EU’s Directive 2010/63/EU, its equivalent of the Animal Welfare Act, emphasizes that animals should only be used if there are no other options and if the potential benefits of the research outweigh the animals’ suffering.
This cost-benefit analysis is subjective. For example, a team of immunologists studying cancer in mice would probably say that the potential public health benefits of their work justify harming mice. A team of science policy experts at PETA would say that mice aren’t ours to use and that these experiments often don’t translate to human trials, anyway.
To bridge this ethical divide, research universities and private companies in the UK have signed a Concordat on Openness on Animal Research, pledging to proactively and transparently inform the public about their treatment of lab animals. In the decade since its launch, nine other countries have followed suit. It’s likely not a coincidence that these countries generally have the tightest restrictions on animal use. However, an independent review found that Concordat signatories in the UK are still struggling to be transparent about their animal research practices in the face of potential disapproval.
On top of the slow pace for necessary regulation, stigma obscures the true nature of what happens in these labs. In the late 2000s, the most extreme opponents of animal testing used violence to try to end the practice, sending poisoned razors and death threats to lab heads and, in at least one case, firebombing a neurobiologist’s car. But rather than encourage scientists to reconsider their methods, attacks like these cemented a culture of silence. While physical violence is not representative of activism against animal testing today — which usually centers around investigations, government advocacy, and direct care for animals and has shifted to become more inclusive — the threat of retaliation still haunts animal researchers, some of whom are encouraged by their institutions to hide their connections to animal testing from the public.
Scientists “don’t want to feel like they’re bad people,” said neuroscientist and author Garet Lahvis, who has written about primate research for Vox.
After learning about what lab animals go through, some people will want to find ways to avoid the products of animal testing. This is much easier said than done, however.
Animal testing is pervasive in health care. Many treatments we take for granted today, like anesthesia, flu shots, and allergy medications, went through preclinical trials in animals before reaching us. They are also valuable to your health, so please keep taking your medicine if you need it.We have more power to avoid animal testing elsewhere. Animal testing requirements are generally looser to nonexistent for cosmetics, cleaning supplies, and other household chemicals, so it’s possible to buy “cruelty-free” makeup or laundry detergent.
The legal distinction between “cosmetics” and “drugs” is blurry, though. Essentially, drugs claim to affect the body’s structure or function in some way, while cosmetics are things you apply to your body to change your appearance (like lipstick) or clean yourself (like deodorant — but not soap, which is neither a cosmetic nor a drug, but its own special category). Many products we might think of as cosmetics are, in fact, also drugs, like anti-dandruff shampoo, tinted moisturizer with sunscreen, and other cosmetics that claim to treat some ailment. In the US, all of these items had to be tested on animals until the FDA Modernization Act 2.0 took effect in 2023.
Cruelty-free claims used on product labels are often misleading, and differences in regulation across countries add to the confusion.
For years, the EU, Canada, Mexico, and 16 other countries (including South Korea, for the skincare girlies) have had legislation in place banning animal testing for cosmetics or their ingredients (although last year, the UK changed their policy to allow testing for makeup ingredients again). But testing on final products or their ingredients has never been banned in the US. Even if a company doesn’t test its final product on animals, it may still run animal tests on raw ingredients. And even if those raw ingredients aren’t currently being tested on animals, they probably were when they were first introduced.
The US government doesn’t have a legal definition for the terms “cruelty-free” or “not tested on animals.” A product labeled “cruelty-free” likely earned voluntary certification from a private organization like Leaping Bunny or PETA’s Beauty Without Bunnies program by pledging to end animal testing at all stages of product development. The definition of “cruelty-free” isn’t standardized across animal protection groups, but earning a “bunny label” generally means that a brand attested to never conducting tests on animals during a product’s development.
Despite pressure from advocates and consumers, many US companies don’t bother with these pledges on animal testing. As of this year, approximately 310 brands globally still test their beauty and household cleaning products on animals. And some actively say they don’t test on animals at all but still sell their products to countries like China, which, until recently, required that all cosmetics (even imported ones) be tested on animals. Most certification programs exclude brands and products sold in China for this very reason.
To make it easier for US companies to sell truly cruelty-free products in China, US regulators and animal welfare advocates have been lobbying their Chinese counterparts for years to change their approach to animal testing for consumer products. Twenty years ago, Thomas Hartung, a toxicologist at the Johns Hopkins Center for Alternatives to Animal Testing, spoke with the National Medical Products Administration (China’s FDA) about regulating animal testing of chemicals and told me “it was like we were coming from Mars.”
In response to yearslong campaigns by organizations like PETA and the Institute for In Vitro Sciences, China recently lifted this requirement. It is now possible to buy Chinese cosmetics that weren’t tested on animals — kind of.
As of January 2021, China no longer requires pre-market or post-market animal testing for cosmetics, meaning that companies from the US and elsewhere can sell things like eyeliner or nail polish in China while still maintaining “cruelty-free” status. But certain “special cosmetics,” like sunscreen, teeth whiteners, and hair dye, or products made for children, are all still required to undergo animal testing. And if a product uses a raw ingredient that isn’t already approved in China, foreign companies have to either reformulate or get that ingredient approved, which requires more animal testing. So, it’s possible to sell US-made “cruelty-free” products in China, but it requires sifting through a confusing and ever-evolving swamp of documentation requirements.
We have made imperfect progress toward a world of cruelty-free cosmetics. While the number of animals used for cosmetic testing in the US has dropped by 90 percent since the 1980s, 44 of the largest 50 cosmetic brands in the world still are not cruelty-free. And without a consensus agreement on what “cruelty-free” actually means, consumers are left to guess which bunny labels are genuine and which are false advertising.
Since many brands can just slap on cruelty-free claims while still sending products abroad to animal testing labs, for now, if you want to avoid animal testing, Leaping Bunny and Beauty Without Bunnies are your best bets. These certifications consider post-market animal testing in other countries as part of their standards.
In some places, like the UK, strict restrictions on animal research and a commitment to transparency have considerably improved lab conditions in recent decades. Companies like Neuralink, however, continue to perform high-risk, ethically dubious experiments hidden from the public eye.
While new alternative methods are under development, animal testing remains necessary in at least some circumstances. Tight regulation — and buy-in from scientists — will be key to minimizing harm in the meantime.
Nicole Kleinstreuer, acting director of the NTP Interagency Center for the Evaluation of Alternative Toxicological Methods (NICEATM), told me that improving the current state of animal testing hinges on researchers gathering “the courage to admit that we can substantially improve upon how we’ve been doing things historically.”
Until relatively recently, alternatives to animal testing in many areas of science were very limited. But in the past decade, bioengineering and computer science have advanced rapidly. New tools like AI, organoids (balls of stem cells that grow into organ tissue), and CRISPR have made replacing animals, at least in certain experiments, more attainable.
For chemical testing, good animal-free research methods have been around for decades — long before most scientists considered using them. Even when well-validated animal alternatives exist, researchers can be slow to adopt them. Hartung, a toxicologist, said, “I turned 60 last year. The methods they’re using were introduced when I was in kindergarten.”
In 2007, the National Academies of Sciences, Engineering, and Medicine, a nonprofit that produces independent policy guidance for the US, laid out a strategy for researchers to move away from using animals in toxicity testing and to develop faster, more human-relevant models to take their place. Today, a number of working groups, both within the US and collaborating internationally, are still trying to put this principle into practice.
As the largest single public funder of biomedical research in the world, the National Institutes of Health (NIH) is uniquely positioned to influence animal testing. In 2023, the NIH spent an estimated $19 billion on US-based projects involving animals, according to Citizens for Alternatives to Animal Research and Experimentation. Between 2011 and 2021, they spent $2.2 billion on projects based in other countries — where oversight boils down to trusting self-generated, non-validated reports from foreign institutions.
Kleinstreuer said that changing the current state of animal research “really necessitates a sea change, and a dramatic investment on the part of funders, particularly the NIH.”
The people in charge of the money have the power to redistribute it and could choose to spend more of it on projects that don’t use animals and less on those that do. That’s the easy part. “It’s kind of the lowest-hanging fruit, and the easiest ask,” said Emily Trunnell, director of Science Advancement and Outreach at PETA. “Even people who are in support of animal testing are on board with the funding of different methods as well.”
NICEATM, led by Kleinstreuer, is doing the in-the-weeds work of figuring out how we’d know whether a replacement method is good enough to substitute for animal experiments. Earlier this year, the NIH also approved the Complement Animal Research in Experimentation (Complement-ARIE) Program, which will set up technology development centers for researchers to make better human-based models.
Non-animal methods can already outperform certain animal tests. Back in 2018, Hartung’s research group created algorithms mapping the relationships between 10,000 known chemical compounds. With this model and lots of data, they predicted the toxicity of 89 percent of the 48,000 toxic chemicals more accurately than animal tests could and for much less money — without endangering any living creatures. Since then, Hartung said things have only become better. But AI-driven research methods are still limited by what real-world data has already been collected. “When you have no data,” he said, “nothing is possible.”
In some cases, using animals is simply bad science. There are some questions “that absolutely necessitate a human cell-based approach,” Kleinstreuer said. “You can’t look at the efficacy of a drug whose target is not expressed in animals by using animal models,” she added. Certain cancer drugs target protein receptors that only exist in humans, and gene therapies often aim to rewrite human-specific DNA sequences. One emerging option: take a sample of human cells, reprogram them to behave like whatever cells you want them to be, and test your drug on the resulting tissue sample.
These tools offer exciting opportunities to personalize medicine to individual patients, but it’s still tough to extrapolate results from a small mass of lab-grown cells in a tightly controlled environment to a human body and the complex interactions of its organ systems. Cancer and embryonic development are incredibly complex biological processes, involving lots of different interconnected body parts that evolve over time. Without that capability, Kleinstreuer said it’s harder to argue that a substance is actually safe and ready to clear for human use.
As it stands, alternatives to animal tests are not being used as widely as they should be, especially in cosmetics. But if we want to study things like deep brain stimulation or run safety tests on new cancer drugs, animal tests are all we have.
While we are stuck with animal experiments, we can try to limit them and make them more humane. Lahvis believes that we should have extremely strict criteria for what animal experiments are funded. Strategically allocating grant funding could not only save millions of lives, but also inspire better science.
Convincing animal researchers to replace animals with other methods is still a huge challenge. Hartung joked that in academia, change happens “one retirement at a time.” Unfortunately, “it’s often been one graveyard at a time,” as retired scientists continue to serve as reviewers who help choose what new projects get funded and published.
The further along a scientist is in their career, the more challenging it becomes to pivot. Because scientists are pushed to maintain a constant level of productivity, Trunnell said, someone who builds their whole lab around their current use of animal models has no incentive to change, unless they have a strong desire to do so. Changing tactics could mean putting their job on the line.
“We’re highly leveraged by the system to keep doing what we’ve always done,” Lahvis agreed. And, Hartung said, turning against a tried-and-true method would require a scientist invalidating their existing body of work or at least acknowledging that it was either unethical, ineffective, or inefficient. Using past observations to inform future experiments is at the core of the scientific method, but, Hartung said, “We’re not trained to be very self-critical.”
That said, a growing number of scientists support the development of non-animal methods, even as they continue to work with animals themselves. People want new tools, whether for the sake of animal welfare or simply because it would make for better science. We might just have to wait another generation.