Research on animals: Is animal testing ok?
Author and Editor: Mariam Kvaratshelia
The Anatomy Lesson of Dr Nicolaes, Rembrandt (1632)
There is a Georgian poem called The Snake-Eater about a man who acquires the ability to communicate with animals after eating a piece of snake.
Once he can understand them, he also understands their pain.
He can no longer eat meat. He tries to convince others to stop eating animals too, but instead of listening, his community rejects him. Eventually, he takes his own life.
The story is tragic, but there is something particularly unsettling about it: before gaining this strange ability, the man simply did not know what animals were saying.
Perhaps we are not so different.
We know that animals are used in laboratories. We know that experiments are performed on them. We know that some experience pain, distress or lasting harm.
But the language we use: research, testing, procedures, trials, models, can make what happens inside a laboratory feel distant and abstract.
So what actually happens when an animal becomes a research subject?
What does “animal testing” really mean?
Animal research is not one single practice.
Animals can be used to study diseases, understand biological processes, develop treatments, test the safety of medicines and chemicals, and conduct other forms of scientific research. The severity and purpose of these procedures vary enormously.
In Great Britain, for example, official statistics recorded 2.54 million scientific procedures involving living animals in 2025. About 1.32 million of these were experimental procedures, while the rest were largely connected to the creation and breeding of genetically altered animals. Mice, fish, birds and rats accounted for the vast majority of experimental procedures.
The statistics also show something that is easy to miss when we simply talk about “animal testing”: not every procedure has the same purpose or severity. In 2025, 54% of experimental procedures in Great Britain were classified as basic research, 26% as applied research and 19% as regulatory testing. Ninety-nine percent were assessed as non-recovery, sub-threshold, mild or moderate in severity, while 1% were classified as severe.
That distinction matters.
But even when a procedure is described as mild, the ethical question does not disappear: when is it acceptable to use another living, sentient being for our own purposes?
The 3Rs: have we found a way forward?
The scientific community has not ignored this question.
In 1959, researchers William Russell and Rex Burch proposed what became known as the 3Rs: replacement, reduction and refinement. More than six decades later, these principles remain central to discussions about responsible animal research.
Replacement means using a non-animal method whenever an appropriate alternative exists.
Reduction means using the smallest number of animals necessary to obtain scientifically valid results.
Refinement means changing procedures or conditions to minimise pain and distress and improve animal welfare.
These principles sound straightforward. But their existence does not mean that animal research has disappeared.
In fact, scientists continue to debate where animal models are genuinely necessary and where they can be replaced.
And this is where the conversation becomes more complicated.
Why do scientists still use animals?
The simplest answer is that animals can provide information that isolated cells or computer simulations cannot always provide.
A living organism is an extraordinarily complicated system. Its organs interact. Its immune system responds. Its hormones change. Its metabolism affects other processes. Some scientific questions therefore require researchers to observe interactions that cannot yet be reproduced completely outside a living body.
Even the U.S. National Institutes of Health, while actively promoting alternatives, states that animal models remain important for some areas of biomedical research because current alternatives cannot reproduce every biological and behavioural aspect of human disease.
There is another problem, however.
An animal is not a small human.
A mouse and a human can share important biological mechanisms, but they are still different species. A result observed in one does not automatically translate into the other.
This is one reason scientists are increasingly interested in methods that model human biology more directly. Research on human-based models such as organoids and organs-on-chips is expanding partly because conventional animal models have limitations when it comes to predicting human responses.
So the question is no longer simply:
“Can we do science without animals?”
It is also:
“Can we develop science that tells us more about humans without first passing through another species?”
What could replace animals?
The alternatives are no longer limited to a scientist staring at a computer screen.
Researchers can work with human cells and tissues, three-dimensional cultures, organoids, mathematical models, computational simulations and microphysiological systems such as organs-on-chips.
Organoids are tiny, three-dimensional structures grown from cells that can reproduce some features of human organs. Organs-on-chips use living cells in engineered systems designed to mimic aspects of how human tissues and organs function.
The U.S. National Institute of Environmental Health Sciences lists these approaches among the growing group of so-called New Approach Methodologies, alongside computational and biochemical methods. Some have already shown the ability to perform as well as, or in certain applications better than, traditional animal models.
But these technologies are not a magic replacement for every experiment.
They have their own limitations, and scientists are still working to make them more reliable, reproducible and capable of representing complex biological systems.
That means the transition away from animal research, where possible, is not simply a matter of banning animals and replacing them overnight.
It is a scientific challenge in itself.
And then there is the ethical question
Science asks what we can do.Ethics asks what we should do.
Humans have built a world around human needs. We decide which species are food, which are companions, which are pests and which become research subjects.
Perhaps this is partly why it is so easy to overlook the individual animal.
A mouse in a laboratory is not usually perceived in the same way as a dog sleeping beside us at home. A fish in a research facility is rarely given the same emotional consideration as a fish in an aquarium.
Yet the fact that an animal lives outside our homes does not make its experience of the world irrelevant.
Animals can experience pain and distress. Modern animal-welfare frameworks recognise this explicitly, and scientific regulations increasingly require researchers to consider and minimise discomfort, distress and pain where animals are used.
The question, then, is not whether every experiment involving an animal is equally harmful.
It is whether the potential human benefit is enough to justify imposing a cost on another living being, and whether that cost could instead be avoided.
Perhaps we don't need to be Snake-Eaters
The man in The Snake-Eater needed a supernatural ability before he could understand that animals suffered.
We don't.
We already know that they respond to pain. We already know that they can experience distress. And we increasingly have technologies that allow us to investigate biology without relying on animals in every situation.
None of this makes the ethical question easy.
Animal research has contributed to scientific knowledge, and some researchers argue that certain experiments still cannot be adequately replaced. At the same time, the development of human-based methods is opening new possibilities for reducing and eventually replacing animal use in areas where alternatives can provide scientifically valid answers.
Perhaps the point is not to imagine a world in which humans suddenly stop caring about human health.
It is to imagine one in which advancing human health no longer automatically means accepting another animal's suffering.
We don't need to understand animals' language to recognise that they have a world of their own.
Finding the balance means taking both sides of the equation seriously: the potential benefits of research and the suffering of the animals involved. It means questioning experiments where alternatives exist, investing in methods that can replace them, and applying the highest possible standards of welfare when animals are genuinely necessary.
All we need is to remember that scientific progress and compassion do not have to be opposites.
Sources & further readingRussell, W. M. S. & Burch, R. L. (1959). The Principles of Humane Experimental Technique. — The original work introducing the principles of Replacement, Reduction and Refinement (the “3Rs”) in animal research.
UK Home Office (2026). Scientific procedures on living animals, Great Britain: 2025. — Official statistics on the number, species, purposes and severity of scientific procedures involving living animals in Great Britain.
National Institutes of Health (NIH). When Are Alternatives to Animals Used in Research? — Overview of emerging alternatives including cell and tissue cultures, 3D tissue models, organs-on-chips and computational models.
National Institutes of Health (NIH). Why Animals Are Used in Research. — Background on why animal models continue to be used in biomedical research and the limitations of current alternatives.
Nature Reviews Bioengineering / Nature. Research and reviews on human-based models, organoids, organs-on-chips and the challenges of translating findings from animal models to humans.
