A student finds the answer key before a test.
Nobody knows.
Using it would probably raise the mark.
Should they?
A tutor discovers that rewriting one paragraph would make a student’s assignment dramatically better.
The student is struggling.
Should the tutor rewrite it?
A school can collect more learner data.
The data could improve personalisation.
It could also become intrusive.
Should the school collect it?
An AI system can finish the work instantly.
The learner has a deadline.
Should they use it?
These are not primarily knowledge questions.
They are not solved by knowing more vocabulary.
They are not solved by calculating faster.
They are not solved simply by asking:
What works?
They require another question:
What should be done, and why?
That is ethical reasoning.
A useful Wintour House definition is:
Ethical reasoning is the disciplined examination of what ought to be done when actions affect people, rights, responsibilities, fairness, welfare, trust or shared rules—by making the facts, stakeholders, values, duties, consequences and trade-offs explicit enough to justify a decision and revise it when relevant information changes.
The phrase justify a decision matters.
Ethical reasoning is not merely having a strong feeling.
Nor is it merely knowing a rule.
Nor is it proving that everyone must share one philosophical theory.
A real ethical decision often contains several legitimate considerations at once.
Honesty.
Care.
Fairness.
Autonomy.
Safety.
Loyalty.
Privacy.
Responsibility.
Benefit.
Harm.
Sometimes they align.
Sometimes they collide.
That is where reasoning matters.
A 2026 systematic review and meta-analysis in Nature Human Behaviour synthesised 92 randomized controlled trials across primary, secondary and tertiary education, involving 12,693 students. Across intervention-versus-control comparisons, ethics and moral-education interventions improved measured moral outcomes on average, with an overall effect of g = 0.65, while the authors also emphasised substantial limitations in study quality and noted that discussion-rich approaches tended to outperform passive information transfer.
That is a useful educational message.
Ethics is not necessarily improved by telling students what to think.
It can be strengthened by giving them serious decisions to reason through.
The Wintour House question is therefore:
If a learner became excellent at ten ethical-reasoning operations, which ten would still matter when school rules, technology, professions and social contexts changed?
Before the Top 10: “I Feel This Is Wrong” Is a Signal, Not Yet an Argument
Moral intuition matters.
A decision may feel unfair.
Cruel.
Dishonest.
Exploitative.
That feeling can be an important warning.
But ethical reasoning asks us to open it.
Why unfair?
Who bears the cost?
Which expectation was violated?
What right or duty is involved?
Would the judgement change if the roles were reversed?
What consequence are we trying to prevent?
A strong ethical thinker does not suppress emotion.
They translate it into inspectable reasons.
Likewise, logic without human meaning is not enough.
A perfectly consistent argument can still depend on a terrible premise.
Ethical reasoning therefore coordinates:
facts,
values,
human impact,
rules,
principles,
context.
The 2025 systematic review of moral reasoning in education reviewed 60 articles and notes the continuing tension among rationalist, intuitionist and integrative approaches, while highlighting case discussion, reflective practice and open dialogue as recurring educational strategies.
The Wintour position is deliberately practical:
notice the intuition, then make the reason visible.
1. Learn to Define the Ethical Decision, Not Merely the Topic
“AI ethics.”
Too broad.
“Privacy.”
Too broad.
“Academic honesty.”
Still broad.
Ethical reasoning needs a decision.
Should a student use AI to generate the final submitted paragraph?
Should a school retain behavioural data for five years?
Should a tutor correct a student’s assessed work after the student has finished?
Now we can reason.
A vague ethical topic invites slogans.
A decision invites analysis.
Use the form:
Should actor A do action B under conditions C?
That creates:
actor,
action,
context.
Then ask:
What is at stake?
What options actually exist?
This resembles Problem Framing, but the owner is different.
Problem Framing identifies the problem.
Ethical Reasoning identifies the normative decision inside it.
Worth learning because: ethical discussion becomes useful only when the learner knows which action, by which actor, under which conditions is being judged.
2. Learn to Separate Facts From Values
“This technology records every student interaction.”
Fact claim.
“That is too invasive.”
Value judgement.
“Students should be able to opt out.”
Normative claim.
These are different sentences.
They can support one another.
They should not be blurred.
Ethical reasoning needs accurate facts.
How much data?
Who has access?
How long retained?
What benefit?
What risk?
But facts alone do not determine the value judgement automatically.
Ten people prefer one option.
Does that make it ethically right?
Not necessarily.
A policy is legal.
Does that make it fair?
Not automatically.
Likewise, values cannot rescue false facts.
A learner may care deeply about fairness while reasoning from incorrect information.
So ask:
What is true?
Then:
Why does it matter ethically?
Critical Thinking retains the factual-evaluative machinery.
Ethical Reasoning keeps the normative bridge.
Worth learning because: good ethical judgement requires both reliable factual premises and explicit values rather than allowing either one to masquerade as the other.
3. Learn to Identify Everyone Who Can Be Meaningfully Affected
A decision may look private.
Often it is not.
Student copies homework.
Affected:
student,
teacher,
classmates if grading is comparative,
future learner-self,
perhaps group partners.
Tutor sends a student’s work to an AI service.
Affected:
student,
family,
tutor,
platform provider,
perhaps the school.
Ethical reasoning therefore asks:
Who has something at stake?
Stakeholder mapping is not a vote-counting exercise.
One vulnerable person does not become irrelevant because ten powerful people prefer another outcome.
The point is to avoid invisible impact.
Ask:
Who benefits?
Who bears risk?
Who has less power?
Who gave consent?
Who could be excluded?
Who inherits the long-term consequence?
This connects to Perspective-Taking.
Top 10 Perspective-Taking Skills Worth Learning asks the learner to reconstruct another person’s information state and constraints.
Ethical Reasoning asks:
What ethical significance do those interests and vulnerabilities have?
Worth learning because: ethical decisions become distorted when people affected by the action are absent from the reasoning simply because they are not present in the room.
4. Learn to Identify Duties, Rights, Rules and Promises
Some ethical considerations are not about maximising outcomes.
They concern obligations.
Do not deceive.
Keep promises.
Protect confidentiality.
Respect consent.
Do not submit someone else’s work as your own.
Rules can encode these duties.
But ethical reasoning should ask what the rule protects.
A school rule against plagiarism protects:
authorship,
fair assessment,
evidence of learning,
trust.
Understanding the purpose helps when the exact case is new.
Rules also have limits.
A rule may conflict with another duty.
Confidentiality may be limited when serious safety concerns arise.
A promise may have been obtained coercively.
So the skill is not:
obey every rule mechanically.
It is:
identify the duty, right or legitimate expectation at stake and understand its scope.
Worth learning because: some actions are ethically significant because they violate obligations or rights even when breaking the rule might produce an attractive short-term outcome.
5. Learn to Trace Consequences Across People and Time
Ethics also asks:
what happens if we do this?
Immediate benefit.
Delayed harm.
Individual gain.
Group cost.
Visible effect.
Hidden effect.
Suppose AI completes a student’s assignment.
Immediate consequence:
deadline met.
Possible delayed consequence:
teacher receives false evidence of capability.
Student loses practice.
Future task becomes harder.
Trust may weaken.
Consequences should therefore be traced beyond the first effect.
This connects to Systems Thinking.
Systems Thinking models dynamic consequences.
Ethical Reasoning asks:
Which of those consequences matter morally, to whom, and with what weight?
Learners should examine:
benefit,
harm,
likelihood,
magnitude,
distribution,
reversibility.
The goal is not to calculate morality like a spreadsheet.
It is to stop consequence claims from remaining vague.
Worth learning because: ethical choices often look different once short-term benefits, delayed harms and the distribution of consequences across people are made visible.
6. Learn to Test Fairness and Consistency
Would I accept the same rule if I were in the other position?
Would I judge the same act differently if my friend did it?
Does the rule apply consistently?
Are similar cases treated similarly?
These are fairness tests.
Consider:
One student receives an exception.
Why?
Relevant difference?
Medical need?
Access requirement?
Or favouritism?
Fairness is not always identical treatment.
Sometimes equitable treatment requires responding to relevant differences.
The reasoning job is:
Which differences are ethically relevant to the decision?
A learner should therefore avoid two shortcuts:
“Everyone must receive exactly the same.”
and
“Every case is different, so consistency does not matter.”
Fairness needs principled distinction.
Worth learning because: consistency tests expose double standards, while fairness reasoning helps distinguish legitimate accommodation from arbitrary unequal treatment.
7. Learn to Recognise Value Conflicts Instead of Pretending One Principle Solves Everything
Honesty matters.
Kindness matters.
Privacy matters.
Safety matters.
Autonomy matters.
Sometimes they collide.
A student reveals a friend is at serious risk but asks for secrecy.
Loyalty and privacy pull one way.
Safety pulls another.
A teacher wants to personalise instruction.
Better support may require data.
Privacy limits data collection.
Ethical reasoning becomes mature when the learner can say:
This is not good versus bad. It is good value versus good value under tension.
Then ask:
Which value has priority here?
Why?
Can we design an option that protects both more effectively?
Can the conflict be reduced rather than simply choosing one side?
This is where simplistic ethics breaks.
Real life contains trade-offs.
Worth learning because: many difficult ethical decisions arise from conflicts among legitimate values rather than from one obviously moral option and one obviously immoral option.
8. Learn to Generate More Than One Ethically Serious Option
People often create a false binary.
Either:
collect all data,
or
collect none.
Either:
ban AI,
or
allow anything.
Either:
tell the secret,
or
ignore the danger.
Ethical reasoning should search for design space.
Could the data be minimised?
Anonymised?
Deleted sooner?
Could AI be allowed for feedback but not final submission?
Could a concern be escalated while sharing only the information necessary?
Better ethical solutions often appear after options are redesigned.
This is where Constraint Reasoning helps.
Top 10 Constraint Reasoning Skills Worth Learning defines the feasible region.
Ethical Reasoning asks whether we can create a feasible option that protects more of the legitimate values simultaneously.
Worth learning because: ethical dilemmas are often worsened by false binaries, and alternative design can reduce conflict among competing values.
9. Learn to Use Reversibility, Publicity and Role-Swap Tests
Some ethical tests are simple and powerful.
Reversibility:
Would I accept this if I were the person affected?
Publicity:
Would I be comfortable explaining this decision openly to the people involved?
Role swap:
Would my reasoning change if my advantage became someone else’s?
These tests do not mechanically produce one answer.
They expose inconsistency.
Suppose a student says:
“It is fine to use hidden AI help because nobody will know.”
Publicity test:
Would you be comfortable telling the teacher exactly how the answer was produced?
If not, why?
That discomfort may reveal a hidden rule or expectation.
The tests are especially useful for learners because they turn abstract ethics into concrete perspective shifts.
They should not replace deeper reasoning.
They are pressure tests.
Worth learning because: reversibility and publicity expose ethical reasons that depend too heavily on secrecy, self-interest or asymmetric treatment.
10. Learn to Decide, Justify and Revisit the Decision After Reality Returns
Ethical reasoning cannot remain endless reflection.
Eventually:
act.
Then explain.
A strong ethical decision states:
the action,
the key facts,
the stakeholders,
the values or duties,
the trade-offs,
the reason for priority,
the remaining uncertainty.
Then reality returns.
Did harm appear?
Was consent misunderstood?
Did the safeguard fail?
Was someone excluded?
Revision may be required.
Ethical responsibility continues after the decision.
This connects to Decision-Making and Verification.
Decision-Making chooses among options.
Verification checks claims.
Ethical Reasoning asks whether the choice remains normatively defensible as its consequences become visible.
Worth learning because: ethical judgement matures when people remain accountable for outcomes and are willing to revise a decision rather than treating initial justification as permanent moral immunity.
The Top 10 Ethical Reasoning Skills as One System
The Wintour House route is:
DECISION → FACTS/VALUES → STAKEHOLDERS → DUTIES/RIGHTS → CONSEQUENCES → FAIRNESS → VALUE CONFLICT → ALTERNATIVE OPTIONS → REVERSIBILITY/PUBLICITY → DECIDE & REVISIT
The quieter version is:
Define what decision is actually being made. Separate what is true from what is valued. Make affected people visible. Identify duties and rights. Trace benefits and harms through time. Test fairness. Admit when legitimate values conflict. Create better options instead of accepting false binaries. Reverse the roles and imagine explaining the choice publicly. Then decide, justify, and remain accountable when reality returns.
That is ethical reasoning.
Not rule memorisation.
Not moral certainty.
Not majority preference.
Not legal compliance alone.
Not kindness alone.
Ethical reasoning is responsibility made inspectable.
Ethical Reasoning Is Not the Same as Critical Thinking
How to Improve Critical Thinking evaluates claims, evidence, assumptions and alternatives broadly.
Ethical Reasoning adds the normative question:
What ought to be done?
Critical Thinking may tell us a factual claim is weak.
Ethical Reasoning decides how facts and values should influence action.
Ethical Reasoning Is Not the Same as Perspective-Taking
Perspective-Taking reconstructs another person’s information, goals and constraints.
Ethical Reasoning asks what obligations or fairness considerations arise from those perspectives.
Understanding someone is not the same as agreeing with them.
But ethical judgement becomes poorer when relevant perspectives are invisible.
Ethical Reasoning Is Not the Same as Empathy
Empathy can make harm vivid.
Important.
But emotion can be uneven.
We may empathise more with people similar to us.
Ethical reasoning needs principles and consistency alongside empathy.
Care matters.
So does fairness.
Ethical Reasoning Is Not the Same as Legality
Legal can be unethical.
Ethical can exceed what law requires.
Law is one constraint and one social institution.
Ethical reasoning asks the wider question.
That does not mean ignoring law.
It means not collapsing “permitted” into “right.”
Ethical Reasoning Is Not the Same as Decision-Making
Decision-Making owns option selection under goals and constraints.
Ethical Reasoning contributes a particular set of evaluative criteria:
rights,
duties,
fairness,
welfare,
trust,
autonomy,
harm.
A decision can be efficient and ethically poor.
Ethical Reasoning Is Not the Same as Argumentation
Top 10 Argumentation Skills Worth Learning owns the construction and testing of defensible positions.
Ethical reasoning can be argued.
But the content of the argument is normative.
Argumentation is the exchange architecture.
Ethical Reasoning supplies the moral decision structure.
For Primary Students
Primary ethical reasoning should be concrete.
A child finds something that is not theirs.
What should they do?
Why?
Someone is excluded from a game.
Fair?
What information would matter?
A rule gives everyone the same thing but one child cannot access it.
Same?
Fair?
Use stories.
Role swaps.
Simple dilemmas.
Ask:
Who is affected?
What would be fair?
What promise or rule matters?
What could happen next?
How would you feel if the roles changed?
Do not demand philosophical vocabulary.
Build ethical structure.
For Secondary Students
Secondary students can begin handling genuine value conflict.
Privacy versus safety.
Freedom versus responsibility.
Loyalty versus honesty.
Individual benefit versus group fairness.
They should learn that ethical reasoning does not mean finding one slogan.
Ask:
Which facts matter?
Which stakeholders?
Which duties?
Which consequences?
What alternative option protects more values?
Secondary school is also where academic-integrity ethics becomes increasingly important.
Authorship.
Collaboration.
AI use.
Citation.
Cheating.
These are not merely rules.
They concern whether assessment evidence is honest.
For JC Students
JC learners should become comfortable with competing ethical frameworks without turning ethics into theory memorisation.
Consequences.
Rights.
Duties.
Fairness.
Character.
Care.
Different lenses may reveal different features of the same decision.
The goal is not to force one universal philosophical winner.
It is to make the reasoning deeper.
JC students can ask:
Which people bear the risk?
Who consents?
Is the benefit fairly distributed?
Would the rule be acceptable if generalised?
Which value is being sacrificed?
Can the policy be redesigned?
This supports mature GP reasoning while remaining non-partisan.
Ethical Reasoning in Science and Research
Science has ethical constraints.
Consent.
Safety.
Data integrity.
Authorship.
Conflicts of interest.
Animal welfare.
Research with vulnerable groups.
A 2024 systematic review of research-integrity teaching distinguishes stages such as ethical awareness, reasoning, motivation and action and notes that evidence about which stages instruction improves remains uneven.
This is an important boundary.
Knowing the rule does not guarantee ethical action.
Ethical reasoning is one component of responsible practice.
Ethical Reasoning in English and GP
GP and literature contain ethical questions everywhere.
What is fair?
What does society owe?
When may autonomy be limited?
Who bears costs?
Which rights conflict?
Strong writing should not merely announce values.
It should expose the trade-off.
Avoid:
“This is unethical because it is wrong.”
Instead:
“This creates a fairness problem because…”
“This violates informed consent because…”
“The benefit is concentrated while the risk is imposed on…”
Now ethics becomes argument rather than moral decoration.
Ethical Reasoning in Studying
Students make ethical learning choices too.
Can I submit AI-generated work?
Can I copy a friend’s answer to understand the method?
Can I share a test paper I was not supposed to distribute?
Can a tutor rewrite this assessed paragraph?
The useful question is not:
Can I get away with it?
It is:
What evidence of my learning is this assessment supposed to produce, and would my action make that evidence false?
That is a powerful educational ethics test.
Ethical Reasoning in the Age of AI
AI creates new dilemmas because capability moves faster than shared norms.
Authorship.
Privacy.
Bias.
Automation.
Dependence.
Transparency.
Accountability.
A 2025 systematic review of AI ethics education found that programmes frequently use case studies and group projects to engage ethical reasoning but also identified weaknesses in rigorous assessment of what students actually learn.
That is instructive.
Talking about AI ethics is not enough.
Students need to reason through real trade-offs.
A strong AI-ethics workflow asks:
Who is affected?
What data entered the system?
Who benefits?
Who bears error?
Can the decision be appealed?
Is the output transparent enough for the stakes?
What human responsibility remains?
Can the task be redesigned to retain the benefits while reducing harm?
AI does not remove ethics.
It multiplies the number of decisions that need it.
The Ethical Reasoning Paradox: A Good Outcome Does Not Automatically Make the Method Ethical
Cheating produces a high mark.
Good outcome for one person.
The method still corrupts assessment evidence.
Consequences matter.
So do duties and trust.
The Ethical Reasoning Paradox: Equal Treatment Can Be Unfair
Give every learner the same staircase.
One uses a wheelchair.
Identical treatment.
Unequal access.
Fairness sometimes requires relevant accommodation.
The Ethical Reasoning Paradox: A Rule Can Protect Ethics and Still Need an Exception
Rules simplify recurring decisions.
But unusual cases exist.
The exception should be justified by the principle the rule was meant to protect—not convenience.
The Ethical Reasoning Paradox: More Transparency Can Create Harm
Transparency is usually valuable.
But exposing private information can violate confidentiality.
Ethical values can conflict.
That is exactly why reasoning is needed.
The Wintour House Test: Does Ethical Reasoning Survive When AI Can Recommend the “Best” Decision?
Yes.
Because “best” depends on values.
AI can estimate consequences.
Summarise policies.
Generate options.
Identify affected groups.
But somebody must still decide:
which harms matter,
which rights cannot be traded casually,
what fairness requires,
who should consent,
what uncertainty is acceptable,
and who remains accountable.
Those are not purely computational outputs.
That is why Ethical Reasoning belongs permanently in the Skills Worth Learning series.
The mature learner can eventually say:
I know what decision is actually being made. I can separate facts from values, identify affected stakeholders, recognise duties and rights, trace consequences, test fairness, acknowledge genuine value conflicts, generate better options, reverse the roles, explain the choice publicly and revisit the decision when its real consequences become visible.
That is ethical reasoning becoming responsible judgement.
Research Anchors
The ten skills above are a Wintour House editorial synthesis, not a claim that moral psychology or education research has validated one universal ten-part taxonomy of ethical reasoning.
The most current broad evidence is the 2026 Nature Human Behaviour systematic review and meta-analysis. It synthesised 92 randomized controlled trials across primary, secondary and tertiary education, involving 12,693 students. In intervention-versus-control comparisons, moral and ethics education improved measured moral outcomes on average (g = 0.65, 95% CI 0.49–0.81). Discussion-rich interventions were more effective than passive information transfer, but the authors emphasised limitations in evidence quality: only one study was rated low risk of bias and preregistration was rare.
The 2025 systematic review of moral reasoning applicable to teacher education synthesised 60 articles and highlights the continuing role of dilemmas, case discussion, reflection and dialogue, while noting strong concentration in particular regions and theoretical traditions.
The 2025 AI ethics education systematic review finds that case studies and collaborative projects are common pedagogical approaches but also identifies a significant assessment gap: many programmes do not use sufficiently strong methods to determine whether ethical understanding or reasoning has actually improved.
A 2024 systematic review of research-integrity teaching examines ethical awareness, reasoning, motivation and action as distinct stages and concludes that claims about instructional effectiveness are unevenly supported across these stages.
The strongest defensible Wintour House conclusion is therefore:
Ethical reasoning is not memorising rules or expressing moral confidence. It is disciplined normative judgement: define the decision, separate facts from values, make affected stakeholders visible, identify rights and duties, trace consequences, test fairness, expose conflicts among legitimate values, generate better options, use reversibility and publicity as pressure tests, and remain accountable enough to revise the decision when new facts or outcomes change the ethical balance.
