Companies test on animals primarily because government regulators require it. Before a new drug, pesticide, or industrial chemical can be sold, agencies like the FDA and EPA demand safety data that, until very recently, could only come from animal studies. Beyond legal mandates, there are scientific reasons: a living body with interconnected organ systems reacts to substances in ways that cells in a dish still cannot fully replicate.
Regulatory Agencies Require It
The most straightforward reason companies test on animals is that the law tells them to. The FDA has historically required animal safety data before any new drug can enter human clinical trials. The EPA requires a battery of animal tests, sometimes called “the six-pack,” before a pesticide can be registered for sale. These six tests cover oral toxicity, skin toxicity, inhalation toxicity, eye irritation, skin irritation, and skin sensitization. Beyond those initial tests, the EPA may also require longer-term studies: subchronic tests that reveal what happens with repeated exposure over weeks, chronic feeding studies that span most of an animal’s life to detect cancer risk, and reproduction studies to check for effects on fertility and offspring.
For pharmaceuticals, the logic is similar but the stakes are even higher. A drug that will be swallowed, injected, or absorbed into the bloodstream every day for years needs extensive safety screening. Regulators want to know which organs are affected, at what dose, and whether harm builds up over time. Historically, the only way to generate that data in a format regulators would accept was through animal studies.
Living Bodies Behave Differently Than Cells in a Lab
The scientific case for animal testing rests on biological complexity. When you swallow a medication, your liver metabolizes it, your kidneys filter it, your immune system reacts to it, and your cardiovascular system distributes it to every tissue. These organ systems talk to each other constantly. A drug that looks safe when tested on liver cells in a petri dish might cause heart damage once it circulates through a living body, or it might produce a toxic byproduct only after the liver breaks it down.
Animal models share extensive similarities with humans in anatomy, physiology, response to injury, and pharmacology. The cardiovascular system, in particular, is well conserved across mammalian species, which is why heart safety testing in animals has been a regulatory staple for decades. These similarities provide a logical rationale for using animals as stand-ins for humans when deliberate human exposure would be unethical.
Chronic effects present a particular challenge. Some toxic reactions only appear after long exposure or in large populations. These are the kinds of adverse events that typically go undetected until data from large numbers of patients over extended periods are analyzed. Animal studies that last months or years are designed to catch these slow-developing problems before a product reaches the public.
The Track Record Is Worse Than Most People Think
Here is the uncomfortable tension at the center of this issue: animal testing is legally required, yet its predictive accuracy is poor. The FDA estimated in 2004 that 92 percent of drugs passing animal tests fail when they reach human trials. More recent analyses suggest that number has climbed to roughly 96 percent. The main causes of failure are lack of effectiveness and safety problems that animal tests did not predict.
This does not mean animal testing catches nothing. It filters out the most obviously dangerous compounds before they ever reach a human volunteer. But the gap between “safe in a mouse” and “safe in a person” is enormous, and it is one of the key reasons scientists and lawmakers have pushed for alternatives.
Cosmetics Are a Different Story
Not all animal testing is legally required. For cosmetics, the picture has shifted dramatically, especially in Europe. The EU banned animal testing on finished cosmetic products in September 2004. Testing on cosmetic ingredients followed in March 2009. A marketing ban, which prevents the sale of any cosmetic tested on animals regardless of where the testing occurred, took full effect in March 2013. This means no company can sell a new cosmetic product in the EU if any ingredient was tested on animals after those dates.
In countries without such bans, cosmetic companies that still test on animals generally do so either because they sell in markets like China (which until recently required animal testing for imported cosmetics) or because they want safety data that goes beyond what non-animal methods can provide. Companies that want to signal they avoid animal testing entirely can seek certification through programs like Leaping Bunny, which requires not just that the company itself avoids testing but that every supplier and third-party manufacturer in its supply chain signs compliance declarations. Independent audits can be required to verify the system works.
Alternatives Exist but Have Real Limits
Cell-based assays, miniature organs grown from human stem cells (organoids), and organs-on-chips, which are small devices lined with living human cells that mimic how an organ functions, have all advanced rapidly. They can model specific organs with impressive accuracy. But they share a fundamental limitation: they cannot replicate the full-body complexity of interconnected systems. As one researcher put it bluntly, “There is no in vitro model that can replace all animal models.” Many lab-based models are still too simplified to predict what happens inside a living body.
The immune system is a particularly tough problem. Your immune response involves dozens of cell types communicating across every tissue in your body. Organs-on-chips can model a liver or a lung, but modeling a liver that is simultaneously being patrolled by immune cells, supplied by a circulatory system, and influenced by hormones from distant glands remains beyond current technology.
The Law Is Starting to Change
A major legal shift arrived in December 2022 when President Biden signed the FDA Modernization Act 2.0. This law removed the longstanding requirement that drugs must be tested on animals before entering human trials. It does not ban animal testing, but it allows companies to use alternatives, including cell-based assays, organoids, organs-on-chips, and AI-based prediction methods, to generate the preclinical safety data the FDA needs.
This is significant because it breaks the regulatory bottleneck. Previously, even if a company had excellent non-animal data, the FDA could still require animal studies. Now, companies can make the scientific case that their alternative methods are sufficient. The legislation reflects a growing recognition that animal models have frequently resulted in therapeutic dead ends, and that newer methods may more accurately predict human responses.
In practice, the transition will be gradual. Regulators still need to be convinced, case by case, that an alternative method is reliable enough for a given type of drug or exposure. For complex systemic effects, whole-body animal models will likely remain in use for years. But for targeted questions, like whether a compound is toxic to liver cells, human-cell-based methods are already more relevant than animal data because they use actual human biology.
Why Some Companies Choose to Test and Others Don’t
The decision to test on animals is not always a single choice made by a company’s leadership. A pharmaceutical company developing a cancer drug has, until very recently, had no legal option but to conduct animal studies. A cosmetics company selling only in the EU has no legal option to conduct them. The difference comes down to what the product is, where it is sold, and what regulators in those markets demand.
Companies in the chemical and pesticide industries face some of the broadest testing requirements. The EPA needs data on effects in birds, fish, aquatic invertebrates, and plants in addition to mammals. These ecological safety tests exist to protect not just human health but entire ecosystems, and no lab-based alternative can yet model a pesticide’s impact on a pond food chain.
For consumer goods companies that operate across multiple countries with conflicting regulations, the calculus gets complicated. A product sold in the EU must not have been tested on animals; the same product sold in certain other markets may need animal data for import approval. Some companies resolve this by maintaining separate product lines. Others withdraw from markets that require animal testing. The Leaping Bunny standard takes the strictest possible position: a certified company cannot allow animal testing for regulatory submission in any country, period.