Ethical Philosophy Guide

Framework Comparison

Applying Frameworks to a Case

Framework application template:

CASE: A hospital must allocate a scarce drug among three patients.

UTILITARIAN ANALYSIS:
- Calculate expected QALYs (quality-adjusted life years) for each patient
- Allocate to the patient who gains the most QALYs
- Consider: indirect effects on family, society, precedent
- Verdict: Maximize aggregate health outcomes

KANTIAN ANALYSIS:
- Can the allocation principle be universalized?
- Does the procedure respect each patient's dignity?
- Are patients treated as ends in themselves, never merely as means?
- Verdict: Fair lottery or first-come-first-served (equal respect)

VIRTUE ETHICS ANALYSIS:
- What would a compassionate, just, courageous physician do?
- Does the decision reflect practical wisdom (phronesis)?
- How does this decision shape the character of the institution?
- Verdict: Context-dependent judgment by virtuous practitioner

CARE ETHICS ANALYSIS:
- What are the relationships and dependencies involved?
- Who is most vulnerable?
- How can we maintain trust and care in the allocation process?
- Verdict: Attend to the most vulnerable, preserve relationships

Constructing Ethical Arguments

Argument Structure in Applied Ethics

Standard philosophical argument form:

THESIS: Autonomous weapons should be banned under international law.

ARGUMENT:
P1: Weapons that cannot reliably distinguish combatants from civilians
    violate International Humanitarian Law (IHL).
P2: Current autonomous weapons systems cannot reliably make this
    distinction in complex combat environments.
P3: There is no clear technological pathway to solving the distinction
    problem within a meaningful timeframe.
C1: Therefore, autonomous weapons violate IHL. (from P1, P2)
C2: Therefore, they should be banned until the distinction problem
    is solved. (from C1, P3, plus the precautionary principle)

OBJECTION 1: AI targeting may be MORE accurate than human soldiers.
RESPONSE: Accuracy in controlled tests does not equal reliability
in unpredictable combat environments. Moreover, IHL requires not
just accuracy but accountability -- someone must be responsible
for each targeting decision.

OBJECTION 2: A ban would disadvantage compliant nations.
RESPONSE: This is an argument for multilateral treaties, not
against the ban itself. The same logic applies to chemical weapons.

The Method of Reflective Equilibrium

Rawls's reflective equilibrium (standard method in applied ethics):

1. Start with CONSIDERED JUDGMENTS
   - Strong moral intuitions about specific cases
   - Example: "Torturing children for fun is wrong"

2. Formulate MORAL PRINCIPLES
   - Abstract rules that explain and organize judgments
   - Example: "It is wrong to cause suffering without justification"

3. Check for COHERENCE
   - Do principles match judgments?
   - Do principles conflict with each other?

4. ADJUST until equilibrium
   - Revise judgments that conflict with well-supported principles
   - Revise principles that conflict with firm judgments
   - Iterate until coherent

5. Result: A REFLECTIVE EQUILIBRIUM
   - Mutually supporting set of principles and judgments
   - Not infallible, but rationally justified

Major Subfields

Bioethics

AI Ethics

Key areas of AI ethics research:

1. FAIRNESS AND BIAS
   - Algorithmic discrimination (hiring, lending, criminal justice)
   - Definitions of fairness (demographic parity, equalized odds, calibration)
   - Impossibility results: some fairness criteria are mutually incompatible

2. TRANSPARENCY AND EXPLAINABILITY
   - Right to explanation (GDPR Article 22)
   - Interpretable vs. post-hoc explanation
   - Epistemic opacity of deep learning

3. ACCOUNTABILITY AND RESPONSIBILITY
   - Responsibility gap: who is liable when AI causes harm?
   - Distributed responsibility across developers, deployers, users
   - Meaningful human control over automated decisions

4. PRIVACY AND SURVEILLANCE
   - Data collection and consent
   - Facial recognition and civil liberties
   - Differential privacy and data protection

5. EXISTENTIAL AND LONG-TERM RISK
   - Alignment problem: ensuring AI pursues intended goals
   - Concentration of power
   - Autonomous weapons and escalation risks

Research Ethics

The Belmont Principles (1979):

1. RESPECT FOR PERSONS
   - Individuals are autonomous agents
   - Persons with diminished autonomy deserve protection
   - Application: Informed consent procedures

2. BENEFICENCE
   - Do no harm (non-maleficence)
   - Maximize benefits, minimize risks
   - Application: Risk-benefit assessment

3. JUSTICE
   - Fair distribution of research burdens and benefits
   - No exploitation of vulnerable populations
   - Application: Equitable subject selection

Modern additions:
- Community engagement and benefit sharing
- Data sovereignty (especially for Indigenous communities)
- Dual-use research oversight

Writing Ethics Papers

Applied ethics paper structure:

1. INTRODUCTION (1-2 pages)
   - Motivating case or real-world problem
   - Thesis statement (clear normative claim)
   - Roadmap of the argument

2. BACKGROUND (2-3 pages)
   - Empirical context (what is the technology/practice?)
   - Existing policy and legal landscape
   - Previous philosophical literature

3. ARGUMENT (5-10 pages)
   - Present the main argument step by step
   - Each premise defended with evidence or reasoning
   - Consider and respond to objections (minimum 2-3)
   - Distinguish your position from nearby views

4. IMPLICATIONS (2-3 pages)
   - Policy recommendations
   - Institutional design proposals
   - Limitations and open questions

5. CONCLUSION (1 page)
   - Restate thesis and main contribution
   - Acknowledge what remains unresolved

Style: Clear, jargon-minimal prose. Philosophy values precision over complexity.

Best Practices

• Engage with the strongest version of opposing views. Straw man arguments are immediately detected by reviewers.
• Ground abstract arguments in concrete cases. Applied ethics must be applied.
• Collaborate with domain experts. Ethics of AI requires understanding AI; ethics of medicine requires understanding clinical practice.
• Distinguish empirical from normative claims. "X is the case" versus "X ought to be the case."
• Check the Stanford Encyclopedia of Philosophy (SEP) for authoritative surveys of any topic before writing.
• Be explicit about your framework. State which ethical tradition you are working within and why.

References

• Stanford Encyclopedia of Philosophy -- Peer-reviewed reference for all areas of philosophy
• Beauchamp, T. L. & Childress, J. F. (2019). Principles of Biomedical Ethics, 8th ed. Oxford UP.
• Floridi, L. & Cowls, J. (2019). A Unified Framework of Five Principles for AI in Society. Harvard Data Science Review.
• Singer, P. (2011). Practical Ethics, 3rd ed. Cambridge UP.
• PhilPapers -- Comprehensive index of philosophy research