Natural Sciences

Ethics

Sean Coleman

2026-05-09

Who Is Responsible?

Science produces knowledge. Knowledge can be used for harm.

• Who bears responsibility?
• The scientist who discovered the principle?
• The engineer who built the application?
• The government that deployed it?
• The user who operated it?

“Now I am become Death, the destroyer of worlds.” — J. Robert Oppenheimer, quoting the Bhagavad Gita, Trinity test, 1945

Open with the Oppenheimer quote — let it land. Don’t immediately explain it. Ask: “Who is responsible here?” This sets up the stakeholder framework that organises the whole session.

Today’s Questions

1. Who are the knowledge stakeholders in science — and what does each owe?
2. What are the ethical constraints on methods? What can science not do to people?
3. What is epistemic harm — and why does it matter?
4. Is there an ethics of what to investigate — not just how?
5. How do we distribute responsibility when knowledge is applied?

Five blocks. The last two (constraints on what to investigate; user responsibility) are often overlooked by students who write about ethics in science. Flag this: “Most TOK essays on this topic stop at methods. Today we go further.”

Knowledge Stakeholders

Three Groups, Three Sets of Responsibilities

Group	Role	Example
Pure scientists	Produce knowledge	Biomedical researchers
Applied scientists / technologists	Translate knowledge into technology	Pharmaceutical companies
Users	Deploy technology in practice	Doctors, patients

Your ethical responsibilities in science depend on which group you belong to.

The stakeholder framework is the organising concept for the whole session. Return to it throughout. The key insight: responsibility does not vanish by being passed down the chain. Each group owns the part of the process it controls.

The Chain of Responsibility

Pure knowledge → Application → Use

Each step is a choice — and choices create responsibility.

Oppenheimer and Rotblat both faced the same choice: stay or leave the Manhattan Project.

• Oppenheimer stayed — and spent the rest of his life as an advocate for nuclear peace.
• Rotblat resigned — when Germany’s defeat removed the original justification.

Both responses are defensible. Which do you find more defensible?

This is the central ethical tension of the session. Don’t resolve it — hold it open. The point is that both are serious moral positions, not that one is clearly right.

Constraints on Methods

Two Principles in Tension

Kantian principle

Some actions are absolutely forbidden.

Persons must be treated as ends in themselves — never merely as means.

No scientific benefit can justify using a person as an unwilling experimental subject.

Utilitarian principle

An action is ethical if it maximises overall good.

Some suffering is acceptable if the benefits clearly outweigh it.

This permits dentistry — and clinical trials.

Most ethical codes combine both: absolute prohibitions and conditional permissions.

The tension between Kant and Mill is a recurring theme in applied ethics. Students should not think one “wins.” Most real ethical codes blend them. The interesting cases are the ones where they pull in different directions.

The Key Requirement: Informed Consent

Both principles converge on one requirement:

Subjects must freely agree to participate, with full understanding of what is involved — and must be free to withdraw at any time.

This is the lesson of the Nuremberg Code (1947).

Spend a moment on the historical context. The Nuremberg Code was a direct response to Nazi medical experiments — grotesque violations of both the Kantian principle (people used as mere instruments) and any plausible reading of the utilitarian principle (the experiments served no legitimate scientific purpose that couldn’t be achieved otherwise). The Code was the founding document of research ethics.

The Nuremberg Code

Nazi physicians conducted brutal experiments on concentration camp prisoners.

The Nuremberg Trials (1945–49) prosecuted 23 doctors. 16 were convicted.

The Nuremberg Code (1947) — First principle:

“The voluntary consent of the human subject is absolutely essential.”

The Declaration of Helsinki (WMA, 2022) built on this:

“The health of my patient will be my first consideration.”

No scientific goal, however valuable, can override this.

The Nuremberg Code is the founding document of research ethics. Ask: “Why was it necessary to write this down? Wasn’t it obvious?” The answer is that institutional and social pressure can override individual moral judgment — the Code is a structural safeguard.

Epistemic Harm

Harm to Knowledge — and Its Consequences

Physical harm from bad science is easy to see. But what about harm to knowledge itself?

Epistemic harm: harm caused by poor-quality knowledge — falsified data, inadequately evidenced claims, suppressed results.

Epistemic harm can cascade into physical harm when flawed knowledge is applied.

The concept of epistemic harm bridges the ethics and epistemology of science. It’s the reason why scientific misconduct is not just professionally wrong but morally wrong.

Case Study: The Wakefield Scandal

Andrew Wakefield (1998): published a paper claiming a causal link between the MMR vaccine and autism.

• The evidence was inadequate.
• The paper was later retracted and Wakefield struck off the medical register.
• But the damage was done: vaccination rates fell.
• Measles outbreaks followed — a preventable disease.

One researcher’s epistemic failure contributed to physical harm at population scale.

This is one of the most important case studies in the ethics of science. Ask: “At what point does publishing without sufficient evidence become unethical?” And: “How should journals and peer review have caught this?” The answer to the second question is complex — peer review failed here.

Academic Integrity

Academic integrity means:

• Honest data collection, analysis, and reporting
• Transparency about methods, limitations, and conflicts of interest
• Proper attribution of sources and co-authors
• Resistance to pressure from funders to exaggerate positive results

Scientists funded by pharmaceutical companies can face pressure to suppress negative results or overstate efficacy.

Peer review and replication exist precisely to guard against these pressures — but they are not infallible.

The funding problem is real and documented. Students should know that “publication bias” — the tendency to publish positive results and bury negative ones — is a systemic problem in several fields, including medicine. The peer review system helps but cannot fully correct for it.

What to Investigate?

Ethics Enters Before the Experiment

Students often focus on the ethics of how science is done. But ethics applies to what is investigated too.

• Some topics are arguably off-limits: designing a weapon to maximise civilian suffering violates the Kantian principle with no offsetting benefit.
• Some topics are politically problematic: racial profiling in criminal investigation may be both ethically wrong and epistemically irrelevant.
• Ethics of omission: failing to investigate something you should.

The ethics-of-omission point is subtle but important. Scientists are responsible not only for what they do, but also for what they choose not to do. Marie Curie’s likely radiation-related illness is a case where not investigating hazards had tragic consequences — but the field was too new to know the risks.

What Gets Funded — and What Doesn’t

Commercial pressures shape the research agenda:

• Orphan diseases (rare conditions): small patient populations → small markets → underinvestment in cures
• Treatments for wealthy countries receive more research than treatments for poor ones
• Animal testing: most happens in applied science (cosmetics, pharmaceuticals) — not pure research

The Thalidomide disaster (1950s–60s): a drug approved without adequate animal testing caused severe birth defects.

The lesson: insufficient testing is an ethical failure, not just a scientific one.

The orphan drugs problem is genuinely hard. If it costs €1 billion to develop a drug and only 10,000 people need it, who pays? Students often assume “just give it to them cheaply” — but then ask: “Who pays for the development?” There’s no easy answer. The goal is to help students see the genuine complexity, not to resolve it.

The User’s Responsibility

Users of scientific technologies bear responsibilities too.

Knowledge is needed to use a technology safely.

Responsibility is distributed: if a user ignores clear safety instructions, the producer cannot be held responsible.

This creates obligations in both directions:

• Producers must communicate clearly
• Users must be sufficiently informed

This is a slightly different angle on the stakeholder framework introduced at the start. The user’s responsibility is real — but it depends on the user having access to adequate information. When producers obscure or suppress safety information, the responsibility shifts back.

Discussion Questions

Question A — Stakeholders and Responsibility

“Joseph Rotblat resigned from the Manhattan Project when its original justification disappeared. J. Robert Oppenheimer stayed.

Which response was more ethically defensible — and does your answer depend on which stakeholder group you place each person in?”

Follow-up: Does a pure scientist bear any responsibility for how their knowledge is applied? Or does responsibility pass entirely to the applier?

This is the central question of the session. There is no correct answer. The interesting positions: (1) Oppenheimer had no choice once the state had the knowledge; (2) Rotblat’s resignation was the only consistent principled response; (3) staying and then advocating for peace was more effective than leaving. Each position has real force.

Question B — Epistemic Harm

“Wakefield’s MMR/autism paper caused more harm than many cases of physical negligence in science.

Should epistemic failures — publishing without sufficient evidence, suppressing results, fabricating data — be treated as seriously as physical harm in legal and professional codes?”

Follow-up: Where does the responsibility lie when a journal peer-reviews but fails to catch a fraudulent paper?

The question of legal liability for epistemic harm is a live policy question. In some jurisdictions, deliberate data fabrication is now a criminal offence. The peer review question is hard — reviewers are usually volunteer experts who cannot detect sophisticated fraud.

Question C — What to Investigate

“Most ethical discussion of science focuses on how science is done. But the reading argues that what is investigated is equally important.

Is there a scientific topic that should not be investigated — and how would you argue for that limit?”

Follow-up: Who should make the decision about what science is off-limits — scientists, governments, or the public?

Good examples for discussion: human cloning research, gain-of-function research (enhancing pathogen transmissibility), autonomous weapons systems, deep brain stimulation for behavioural control. The governance question (who decides) is crucial — and has no clean answer.

Exit

Three Things to Leave With

1. Responsibility is distributed, not single. Pure scientists, applied scientists, and users each own a part of the chain. No one can fully offload their responsibility onto another group.

2. Epistemic harm is real harm. The obligation to produce honest, well-evidenced knowledge is an ethical obligation — not just a professional norm. When it fails, physical harm can follow.

3. Ethics applies before the experiment begins. What is investigated — and what is not — is an ethical choice. Omissions matter as much as commissions.

Exit Ticket

Write 2–3 sentences — choose one:

Option 1: “Identify one case from this session where the Kantian principle and the utilitarian principle point in different directions. Which principle do you think should take priority, and why?”

Option 2: “Give an example of epistemic harm from this session or your own knowledge. What mechanism should prevent it — and why might that mechanism fail?”

Option 3: “Oppenheimer or Rotblat: whose response was more ethically defensible? Argue a position in 2–3 sentences.”

Reference

Key Terms

Term	Definition
Knowledge stakeholder	Producer, applier, or user of scientific knowledge — each with different responsibilities
Kantian principle	Persons must be treated as ends in themselves; some actions are absolutely forbidden
Utilitarian principle	An action is ethical if it maximises overall good; some harm is acceptable for sufficient benefit
Informed consent	Free, informed, revocable agreement to participate in research
Nuremberg Code	First formal research ethics code (1947); established voluntary consent as foundational
Epistemic harm	Harm to knowledge quality that can cascade into physical harm
Academic integrity	Honest data, transparent methods, proper attribution, resistance to funding pressure
Knowledge virtue	Epistemic disposition (honesty, carefulness, openness) that makes a knower reliable
Ethics of omission	Responsibility for not investigating something that should have been studied

TOK Connections

• Core Theme: Knowledge and Responsibility — Who bears responsibility when knowledge is misused?
• Knowledge and the Knower — Knowledge virtues apply to every knower, not only scientists
• Human Sciences AoK — Informed consent and epistemic harm are central to psychology and sociology too
• History AoK — How do we know about the Nazi experiments? What are the primary sources?
• Ethics — Kant vs. Mill as frameworks; neither is sufficient alone