Why Are We Still Testing On Animals - Part 3
In This Article
- The Future of Medical Research Is Already Here
- Growing Human Biology in a Laboratory
- Putting Human Organs on a Chip
- Artificial Intelligence Is Becoming a Research Partner
- Human Cells Tell Human Stories
- No Single Method Has to Replace an Entire Animal
- This Is Not an Argument for Reckless Replacement
- We Have Been Asking the Wrong Question
- History Shows That Innovation Often Needs a Push
- Cleaner Air Did Not Happen by Accident
- Fuel Economy Standards Changed Vehicle Design
- The Ozone Layer Did Not Recover Because We Waited
- Healthcare Has Already Experienced Forced Innovation
- What Does This Have to Do with Animal Testing?
- What Needs to Happen Next?
- Congress Opened the Door
- The FDA Should Make the New Path Predictable
- The NIH Should Fund the Future
- Universities Must Change What They Teach
- Research Proposals Should Justify the Model
- Validation Must Be Fair
- Transparency Will Accelerate Progress
- Animal Welfare Must Be Part of Scientific Quality
- Success Must Be Measured by Human Outcomes
- The Question Is No Longer Whether We Can Change
The Future of Medical Research Does Not Have to Depend on Animals
By John Lieberman, MD, MSL
Summary
The alternative to animal testing is not less testing.
It is better, more human-relevant testing.
Researchers can now grow organoids from human cells, recreate organ functions on microchips, analyze patient-specific tissues, use artificial intelligence to predict toxicity, and study patterns in real-world clinical data.
These technologies will not replace every animal study immediately. But they have already ended the argument that animal testing is the only scientifically credible path.
The question now is whether laws, regulators, research funding, and universities will help better methods become the new standard.
What You’ll Learn
- What New Approach Methodologies are
- How organoids and organs-on-chips work
- How artificial intelligence can improve drug research
- Why no single method must replace an entire animal
- How regulation has accelerated innovation in other industries
- What the FDA Modernization Act 2.0 changed
- What the FDA, NIH, universities, and researchers should do next
Estimated reading time: 14 minutes
The Future of Medical Research Is Already Here
If animal testing is not the best tool for every scientific question, what comes next?
The answer is not to stop evaluating new medicines.
It is to evaluate them in ways that are more relevant to human biology.
Researchers now have access to a growing collection of tools known as New Approach Methodologies, or NAMs.
NAMs include laboratory and computational methods used to evaluate the safety, effectiveness, and quality of drugs and other products. They may involve human cells, tissues, microengineered systems, artificial intelligence, chemical testing, computer models, and real-world data.
The FDA is actively developing a framework to reduce, refine, and replace animal use with advanced methods that may better predict how medicines work in people.
These technologies are not all equally mature.
Some already support regulatory decisions.
Others remain experimental.
Some answer narrow questions.
Others combine several types of evidence.
No individual method can reproduce an entire human body.
But that is the wrong standard.
An animal cannot fully reproduce a human body either.
The relevant question is whether a method answers a specific scientific question more reliably than the method it may replace.
Growing Human Biology in a Laboratory
Organoids are small, three-dimensional structures grown from stem cells or other human cells.
They can reproduce selected features of organs such as the brain, liver, kidney, lung, and intestine.
An organoid is not a complete miniature organ.
It may lack blood vessels, immune cells, nerves, or other structures.
But organoids allow human cells to organize and interact in ways that flat cell cultures cannot.
Researchers can expose them to medicines.
They can study disease development.
They can observe toxic effects.
They can compare tissues with different genetic characteristics.
In some cases, scientists can grow organoids from an individual patient’s cells.
That opens the door to a more personalized form of research.
Instead of asking whether a treatment works in a standard laboratory animal, researchers may be able to ask how it affects tissue carrying the patient’s own disease-related characteristics.
Imagine testing several cancer treatments on a model derived from a patient’s tumor before selecting therapy.
Imagine studying a rare genetic disease in tissue created from the person affected by it.
Imagine comparing responses across many human donors rather than relying on one strain of mouse.
These applications are still developing.
But they begin with the biology medicine is intended to treat.
Putting Human Organs on a Chip
Tissue chips, often called organs-on-chips, contain living human cells inside small engineered devices.
Microscopic channels carry fluid through the system.
Mechanical forces may imitate breathing, blood flow, stretching, or other physical conditions.
A lung chip may reproduce aspects of expansion and contraction.
A heart chip may contain contracting cardiac cells.
A liver chip may help researchers examine metabolism and toxicity.
Researchers can connect several tissue systems to study interactions between organs.
The Tissue Chip for Drug Screening Program at the National Center for Advancing Translational Sciences was created to improve predictions about whether drugs will be safe or toxic in humans.
Tissue chips are not complete people.
But they offer something an animal cannot.
They allow scientists to study living human cells in an engineered environment designed around a particular human biological question.
Artificial Intelligence Is Becoming a Research Partner
Drug development generates enormous amounts of data.
Researchers study genes, proteins, chemicals, cells, tissues, medical images, clinical outcomes, and adverse events.
No human team can examine every possible relationship manually.
Artificial intelligence can help researchers identify patterns across those datasets.
AI systems may be used to:
- Predict whether a compound is likely to be toxic
- Identify promising drug targets
- Screen large numbers of possible medicines
- Model interactions between drugs
- Design new molecules
- Match patients with clinical trials
- Find existing medicines that may treat other diseases
- Identify groups of patients likely to respond differently
- Detect safety signals in real-world data
Artificial intelligence does not eliminate the need for laboratory or clinical research.
A prediction is not proof.
The quality of the result depends on the quality of the data and assumptions behind the model.
But AI can help scientists eliminate poor candidates earlier and focus resources on the most promising ones.
That may reduce the need to test large numbers of compounds in animals merely to narrow the field.
Human Cells Tell Human Stories
Human-based research allows scientists to ask more direct questions.
What happens in human heart cells carrying a disease-related mutation?
How do neurons derived from a patient respond to a treatment?
What happens when a medicine passes through a human liver model?
How do immune cells from different people respond?
Why does a treatment help one group of patients but not another?
This shift also helps research confront human diversity.
Patients differ by age, sex, genetics, health history, environment, ancestry, medication use, and many other factors.
Laboratory animals are often bred for genetic similarity and kept under controlled conditions.
That consistency can make an experiment easier to analyze.
It may also make the result less representative of the population that will receive the treatment.
Human cells and data can be drawn from many individuals.
Researchers can study variation instead of treating it as noise.
That may be essential to developing safer and more personalized medicine.
No Single Method Has to Replace an Entire Animal
One of the most misleading objections to NAMs is that a cell culture, organoid, or computer model cannot reproduce a whole living body.
That is true.
It is also unnecessary.
A modern research program does not need one new method to replace every function of an animal.
It can combine several methods.
A computer model may predict how a drug is absorbed.
A liver chip may examine metabolism.
A heart-tissue model may detect cardiac toxicity.
Human immune cells may reveal inflammatory responses.
Organoids may show how a particular tissue reacts.
Clinical data may identify outcomes in real patients receiving related medicines.
Together, these methods may provide a more human-relevant picture than one animal species.
The future is not one-for-one replacement.
It is integrated evidence.
This Is Not an Argument for Reckless Replacement
Some scientific questions remain difficult to answer without examining interactions across an entire living organism.
The circulatory, immune, nervous, endocrine, and metabolic systems influence one another.
No single organoid or chip reproduces all of those relationships.
Some animal research may remain necessary while better methods develop.
A responsible transition should not replace a validated animal test with an unproven method merely to claim progress.
The standard must remain scientific performance.
Replace animal use when a validated method provides equal or better information.
Reduce animal use when fewer animals can answer the question.
Refine remaining procedures to minimize suffering.
Invest aggressively in methods capable of replacing more animal studies.
Animal testing should not receive automatic preference because it is old.
NAMs should not receive automatic preference because they are new.
Every model must earn confidence through evidence.
We Have Been Asking the Wrong Question
For years, the debate has centered on one question:
Can we replace animal testing?
That question assumes animal testing should remain the default until every uncertainty about an alternative has been eliminated.
A better question is:
Which method best predicts the human outcome we need to understand?
Sometimes the answer may still involve an animal.
Increasingly, it may involve human cells, tissues, clinical data, or computation.
The method should follow the question.
The question should not be forced into the method an institution already owns.
New methods should be validated.
Their limitations should be disclosed.
Their results should be reproducible.
Animal models deserve exactly the same scrutiny.
Longevity is not proof of superiority.
Tradition is not validation.
History Shows That Innovation Often Needs a Push
Large institutions rarely transform because everyone suddenly agrees.
They change when incentives, expectations, and standards change.
Economists sometimes discuss this idea through the Porter Hypothesis, which proposes that well-designed regulation can encourage innovation instead of merely adding costs.
The concept is visible throughout modern history.
Cleaner air did not appear because every polluting industry voluntarily changed.
Safer and more efficient cars did not emerge because manufacturers independently agreed to absorb every cost.
The ozone layer did not begin recovering because companies voluntarily abandoned profitable chemicals.
Electronic health records did not spread because every hospital eagerly replaced paper.
Policy established a direction.
Innovation responded.
Cleaner Air Did Not Happen by Accident
By the late 1960s, air pollution was impossible to ignore.
Smog covered major cities.
Industrial emissions and automobile exhaust threatened human health.
Congress responded with the Clean Air Act of 1970 and later amendments.
Industries warned that stronger standards would be expensive and technically difficult.
Engineers developed cleaner engines, catalytic converters, fuel injection, improved emission controls, and more sophisticated engine-management systems.
The regulation did not tell every engineer exactly what to invent.
It established performance expectations.
Innovation found ways to meet them.
Fuel Economy Standards Changed Vehicle Design
Congress established the Corporate Average Fuel Economy Program during the 1970s.
The standards required automobile manufacturers to improve fleet fuel efficiency.
Companies responded through lighter materials, improved aerodynamics, more efficient engines, better transmissions, hybrid systems, and other innovations.
Again, the government established the destination.
Engineers chose the route.
The Ozone Layer Did Not Recover Because We Waited
Scientists discovered that chlorofluorocarbons were damaging Earth’s protective ozone layer.
Those chemicals were widely used in refrigeration, air conditioning, foam products, and aerosols.
Replacing them appeared expensive and disruptive.
Countries adopted the Montreal Protocol on Substances That Deplete the Ozone Layer in 1987.
The agreement created schedules for reducing and eliminating damaging chemicals.
Industry developed alternatives.
Governments adjusted the policy as science evolved.
The world did not wait until every replacement technology was perfect.
It established a direction and timetable.
Innovation continued during the transition.
Healthcare Has Already Experienced Forced Innovation
Hospitals and medical practices once stored almost every patient record on paper.
Records could be difficult to find, read, or share.
Many organizations delayed switching to electronic systems because the transition was expensive and disruptive.
The Health Information Technology for Economic and Clinical Health Act, known as the HITECH Act, changed the incentives.
The law supported the adoption and meaningful use of electronic health records.
The transition was not painless.
Electronic systems created new frustrations and administrative burdens.
Yet the policy moved healthcare away from a system that many institutions would otherwise have maintained much longer.
The lesson is not that every regulation produces perfect results.
The lesson is that coordinated policy can move a resistant system.
What Does This Have to Do with Animal Testing?
Everything.
Animal testing remains dominant partly because the system rewards it.
Universities maintain animal facilities.
Researchers are trained in animal methods.
Grant programs support familiar models.
Regulators understand traditional evidence.
Companies know how to produce it.
Human-based methods must compete against infrastructure built over many decades.
Waiting for the system to transform itself may mean waiting indefinitely.
Government does not need to dictate which chip, organoid, or computer model every scientist must use.
It can establish the goal.
It can require stronger scientific justification for animal experiments.
It can create clear validation pathways for human-based methods.
It can shift funding toward research designed around human biology.
It can establish timelines for replacing specific animal tests when qualified alternatives exist.
That is not interference with science.
It is a decision to stop giving old science an automatic advantage.
What Needs to Happen Next?
The United States has already begun moving in this direction.
The question is whether early reforms will produce structural change.
Congress Opened the Door
The FDA Modernization Act 2.0 changed federal law by broadening the definition of acceptable nonclinical testing.
The law recognizes methods that may include cell-based assays, microphysiological systems, bioprinted models, and computer-based approaches.
It did not ban animal testing.
It ended the assumption that animal studies were the only legally recognized path.
That was a historic change.
But permission alone does not transform a regulatory culture.
Companies need confidence that validated NAMs will be accepted.
Researchers need clear guidance.
Regulators need training.
Developers need predictable standards.
The law opened the door.
Implementation determines whether the scientific community walks through it.
The FDA Should Make the New Path Predictable
The FDA is now doing more than acknowledging alternatives.
Its New Approach Methodologies Program promotes human-relevant methods that include artificial intelligence, organs-on-chips, and real-world data.
The agency’s Innovative Science and Technology Approaches for New Drugs Program, known as ISTAND, provides a pathway for emerging drug-development tools.
In March 2026, the FDA issued the draft guidance General Considerations for the Use of New Approach Methodologies in Drug Development.
The draft provides a validation framework and recommendations for submitting NAM data in drug development. It also states that improving human relevance while reducing animal use is an FDA priority.
These steps show that change is no longer theoretical.
The FDA should continue by:
- Publishing clear standards for deciding when a NAM can replace an animal study
- Creating predictable review pathways
- Sharing successful regulatory case studies
- Training reviewers across agency divisions
- Coordinating requirements with international regulators
- Avoiding unnecessary requests for animal data
- Measuring whether NAMs improve safety, cost, speed, and clinical success
Regulatory uncertainty protects the old system.
Clear expectations encourage investment in the new one.
The NIH Should Fund the Future
The National Institutes of Health shapes medical research through its funding decisions.
Researchers pursue important scientific questions.
They also pursue projects likely to receive grants.
When funding favors familiar animal models, those models remain dominant.
When funding supports organoids, tissue chips, computational biology, and human data, expertise grows in those fields.
The NIH Complement Animal Research in Experimentation Program, known as Complement-ARIE, is designed to speed the development, standardization, validation, and use of human-based NAMs.
In March 2026, the NIH announced more than $150 million in investments to develop and scale methods intended to simulate human biology more accurately and reduce reliance on animal models.
That is meaningful progress.
It should be the beginning.
The NIH should also:
- Require stronger justification for proposed animal models
- Fund direct comparisons between animal and human-based methods
- Support replication and validation studies
- Build shared organoid and tissue-chip facilities
- Train researchers in computational and human-based science
- Reward publication of negative results
- Develop databases comparing model predictions with clinical outcomes
- Redirect funding when a model repeatedly fails to translate
A funding system should not preserve a method because researchers have always used it.
It should reward the method most likely to answer the question.
Universities Must Change What They Teach
Research culture begins in training.
Graduate students and young scientists learn which methods senior researchers trust.
They learn what grant reviewers expect.
They learn what journals publish.
They often inherit the models their laboratories already use.
Universities should prepare future scientists to work with:
- Human cells and tissues
- Organoids
- Microphysiological systems
- Computational biology
- Bioinformatics
- Artificial intelligence
- Clinical research design
- Patient-derived data
- Translational science
- Research ethics
- Comparative model evaluation
This does not mean erasing knowledge of animal biology.
It means ending the assumption that an animal experiment is the natural starting point for every biomedical question.
Research Proposals Should Justify the Model
Scientists must justify many aspects of a study.
They explain their sample size.
They identify controls.
They describe statistical methods.
They disclose limitations.
They should also explain why the selected model is the best one available.
A proposal involving animals should answer several questions:
- What exact human process is the animal expected to reproduce?
- What important biological differences exist between the species?
- How well has the model predicted human outcomes before?
- Could a human-based method answer part or all of the question?
- Could several methods be combined?
- How will the findings be verified in human biology?
- Is the expected knowledge sufficient to justify the animal’s experience?
These are not bureaucratic obstacles.
They are scientific questions.
Validation Must Be Fair
New methods are often tested against existing animal results.
That comparison can become circular.
A human-based model may predict human outcomes more accurately but appear unsuccessful because it does not reproduce an animal response.
Animal data may be part of the comparison.
It should not automatically be treated as the truth.
The final benchmark must be human biology.
Does the method predict human toxicity?
Does it reproduce human metabolism?
Does it identify an immune response seen in patients?
Does it help determine a useful dose?
Does it predict clinical effectiveness?
Does it account for differences among people?
A method that predicts human outcomes better should receive priority.
Transparency Will Accelerate Progress
Research institutions, pharmaceutical companies, and regulators hold enormous amounts of data.
Some animal models may perform well for specific purposes.
Others may repeatedly fail.
Researchers cannot easily distinguish them without transparent information.
Organizations should report:
- Which models were used
- What each model predicted
- Whether later human evidence agreed
- Why development ended
- Whether toxicity appeared in animals, humans, or both
- Whether a NAM produced a better prediction
- How much time and money each approach required
Negative results are especially important.
When failed studies remain hidden, other researchers may repeat them.
More animals may be used.
More money may be spent.
Patients may wait longer.
Transparency is not merely an ethical obligation.
It is an innovation strategy.
Animal Welfare Must Be Part of Scientific Quality
Reducing suffering should not be treated as separate from improving science.
Stress, fear, pain, isolation, and abnormal housing can alter biology.
They can affect immune function, hormones, metabolism, behavior, and disease response.
A research system that takes animal experience seriously may also produce more reliable results.
But refinement alone is not enough.
A painless unnecessary experiment would still be unnecessary.
The first question should not be how to make an animal experiment less harmful.
The first question should be whether the animal is needed.
Only after that question is answered should researchers determine how to reduce suffering in any remaining work.
Success Must Be Measured by Human Outcomes
The goal is not to eliminate animal testing merely to achieve a numerical target.
The goal is better medicine.
Safer drugs.
More accurate predictions.
Fewer failed trials.
Faster discoveries.
Lower costs.
Treatments that help patients.
Reducing animal suffering is an essential moral outcome.
It may also be evidence that research has become more precise.
A poor test is not made scientific because it has existed for decades.
A new test is not made reliable because it appears technologically impressive.
Every method must earn confidence through performance.
That standard should apply equally.
The Question Is No Longer Whether We Can Change
Animal research has contributed to scientific knowledge.
It has helped researchers understand physiology, disease, immunity, genetics, and treatment.
Acknowledging those contributions does not require permanent loyalty to the methods that produced them.
Science advances by questioning its assumptions.
It replaces older tools when better ones become available.
It follows evidence even when that evidence challenges established institutions.
Today, researchers can grow human tissues.
They can connect organ systems on chips.
They can analyze enormous datasets with artificial intelligence.
They can study patient-specific biology.
They can combine laboratory evidence with clinical and real-world data.
These methods will not solve every research problem tomorrow.
But they have already made one fact impossible to deny.
Animal testing is no longer the only path.
In many areas, it may no longer be the best path.
Every unnecessary experiment creates unnecessary suffering.
Every year of delay preserves a system that often struggles to predict human outcomes.
Every missed opportunity slows the development of methods designed around the species medicine is supposed to help.
The question is no longer whether alternatives exist.
They do.
The question is whether our laws, funding systems, universities, regulators, and scientific institutions will give them a genuine chance to succeed.
History will not judge us by how faithfully we protected the methods of the past.
It will judge us by whether we had the courage to improve them.
The true purpose of medical research has never been to preserve tradition.
It is to improve human health.
When that goal can be pursued more accurately, efficiently, and humanely, moving forward is not simply an opportunity.
It is a responsibility.
References
- U.S. Food and Drug Administration. New Approach Methodologies.
- U.S. Food and Drug Administration. General Considerations for the Use of New Approach Methodologies in Drug Development. Draft guidance. March 2026.
- National Center for Advancing Translational Sciences. Tissue Chip for Drug Screening.
- FDA Modernization Act 2.0. S.5002, 117th Congress.
- U.S. Food and Drug Administration. Innovative Science and Technology Approaches for New Drugs.
- National Institutes of Health Common Fund. Complement Animal Research in Experimentation Program.
- National Institutes of Health Common Fund. NIH Invests $150 Million in Human-Based Research to Reduce Use of Animal Models. March 18, 2026.
- U.S. Environmental Protection Agency. Clean Air Act Text.
- National Highway Traffic Safety Administration. Corporate Average Fuel Economy.
- United Nations Environment Programme. Montreal Protocol on Substances That Deplete the Ozone Layer.
- Office of the National Coordinator for Health Information Technology. Health IT Legislation.
