02/10 Recognizing and Framing Ethical Problems

Recognizing the Ethical Dimension

Ethical analysis begins by determining what the moral problem actually is, before selecting an ethical theory or defending a decision. A situation has an ethical dimension when it concerns how people ought to act or how people should be treated.

Common indicators include:

• Someone may be harmed, helped, exploited, deceived, excluded, or placed at risk.

• Rights, duties, responsibilities, or legitimate interests may conflict.

• Benefits, burdens, opportunities, or punishments may be distributed unfairly.

• Important values may point toward different actions.

• A decision may affect human dignity, relationships, vulnerability, or character.

An ethical problem is not always a choice between an obviously good option and an obviously bad one. Difficult cases often involve two legitimate goods, such as privacy and public safety, or two serious responsibilities, such as protecting confidentiality and preventing foreseeable harm.

To diagnose an issue, ask:

1. Could someone be harmed, exploited, deceived, unfairly treated, or denied an important opportunity?

2. Does the situation involve conflicting rights, duties, values, or interests?

3. Are some people carrying risks or burdens so that others receive benefits?

4. Would the decision remain acceptable if its effects were publicly explained?

5. Is the issue about more than efficiency, custom, personal preference, or legal compliance?

Takeaway: Moral discomfort can be a useful starting signal, but ethical analysis requires identifying the specific effects, duties, values, and people involved.

Framing a Question for Moral Inquiry

A vague concern becomes useful for analysis when it is converted into an ethical question. The wording should identify the proposed action, the circumstances, the affected parties, and the competing considerations without assuming the answer.

For example, “Should the organization adopt the new monitoring software?” focuses only on adoption. A stronger formulation is:

Should the organization use monitoring software that may improve cybersecurity but collect employees’ personal information without their full understanding or consent?

This version makes privacy, security, consent, and possible unequal effects visible. A useful problem statement can be built from:

Decision-maker + proposed action + relevant circumstances + affected stakeholders + competing values.

A university example might be framed as follows:

A university must decide whether to use automated examination-proctoring software when the system may reduce cheating but may also collect sensitive data, misidentify students, and create unequal burdens for students with disabilities or limited access to private spaces.

A strong framing should be:

• Specific: It identifies the decision and its circumstances.

• Neutral enough for inquiry: It does not assume the conclusion.

• Fact-sensitive: It distinguishes established information from uncertainty.

• Stakeholder-aware: It includes people who may otherwise be overlooked.

• Value-explicit: It names competing ethical considerations.

• Action-guiding: It can lead to practical choices.

Takeaway: A good frame does not settle the issue; it creates a precise subject for justified moral inquiry.

Separating Evidence from Evaluation

The fact–value distinction separates descriptions of conditions from judgments about what ought to be the case. A fact claim reports or explains what happened, what exists, or what is likely. A value or normative claim evaluates something as good, bad, right, wrong, fair, unfair, important, necessary, or permissible.

Examples:

• “The software records video during an examination” is a factual claim that requires accurate documentation.

• “The software has a higher error rate for some students” is a factual claim that requires evidence about the size and significance of the difference.

• “Students should be told exactly what data are collected” is a normative claim that requires justification for transparency or informed consent.

• “It is unfair to impose greater burdens on students who lack private study space” is an evaluative judgment that depends on an account of fairness.

Facts and values are distinct, but ethical decisions require both. Facts help establish which actions are possible and what consequences are likely. Values help assess whether those consequences, procedures, and relationships are acceptable. The claim that an action will increase efficiency does not, by itself, prove that the action ought to be taken.

Organize case information into four lists:

1. Known facts: Information supported by reliable evidence.

2. Unknown facts: Information that could change the decision but has not been established.

3. Assumptions: Beliefs treated as true without adequate support.

4. Value judgments: Claims about what is good, harmful, fair, right, necessary, or permissible.

For instance, “Most students do not object to monitoring” is a factual claim that needs evidence. “Students should accept monitoring because they have nothing to hide” is an argument built on assumptions about privacy and consent. Keeping these categories separate prevents a value judgment from being disguised as a fact.

Takeaway: Accurate evidence does not eliminate the need for moral judgment, and moral judgment should not be presented as though it were evidence.

Testing Facts and Clarifying Concepts

Ethical reasoning depends on investigating the circumstances rather than reacting to an incomplete picture. Before evaluating options, determine what is known, what is uncertain, and what further information could change the analysis.

Relevant questions include:

• What exactly happened, or what action is proposed?

• Who has authority to act?

• What realistic alternatives are available?

• What benefits and harms are likely for each alternative?

• How severe, likely, and reversible are the possible harms?

• Who has information that the decision-maker lacks?

• What legal, professional, institutional, or technical constraints matter?

• Could effects differ because of disability, race, age, income, status, or location?

• What less harmful alternatives have been considered?

A fact is relevant when it could affect the ethical evaluation of an option. For example, the cost of a policy matters ethically if it determines who can access a benefit or whether resources are diverted from a more urgent need.

Important concepts also require clarification. Terms such as harm, privacy, consent, fairness, responsibility, discrimination, necessity, and public interest can support different conclusions when people use them differently. Clarification should make a claim more precise rather than merely replace one vague word with another.

Compare:

• “The policy is unfair.”

• “The policy is unfair because it imposes a substantially greater cost on students who cannot access private testing spaces while providing no effective alternative.”

The second statement identifies the relevant meaning of fairness and the facts needed to assess it.

Takeaway: Investigate only information that could affect the decision, and define central concepts in a way that connects them to observable circumstances.

Mapping Affected People and Interests

A stakeholder is anyone with an important interest in, or exposure to, the outcome of a decision. Stakeholders may experience direct effects, indirect consequences, burdens, benefits, or changes to their rights and responsibilities.

Look beyond the people who make the decision or the most visible group. Consider:

• Direct stakeholders: People who experience immediate effects.

• Indirect stakeholders: People affected through wider social, economic, or institutional consequences.

• Vulnerable stakeholders: People with reduced power, fewer alternatives, or greater exposure to harm.

• Future stakeholders: People who may be affected later, including future generations.

• Voiceless or underrepresented stakeholders: People unable to participate effectively or represent their own interests.

• Institutional stakeholders: Organizations responsible for duties, resources, or public trust.

• Environmental or nonhuman stakeholders: Animals, ecosystems, or other parts of the natural world when affected.

For each stakeholder, ask:

1. What interests, rights, needs, or responsibilities are involved?

2. What benefits and harms might result?

3. How much power does the stakeholder have in the decision?

4. Has the stakeholder been consulted or represented accurately?

5. Are burdens distributed differently among groups?

6. Could this stakeholder’s interests conflict with those of another group?

Identifying a stakeholder does not automatically determine how much weight that person’s interest should receive. It ensures that the interest is visible and must be addressed rather than ignored. Special attention to a vulnerable stakeholder is important because unequal power or limited alternatives can make an apparently neutral policy impose unequal burdens.

Takeaway: Stakeholder analysis expands the moral field so that indirect, future, vulnerable, and underrepresented interests can be considered.

Comparing Options and Ethical Tensions

After gathering facts and identifying stakeholders, map the ethical structure before choosing an action. List more than the two options initially presented. Possible options include:

• taking the proposed action;

• maintaining the status quo;

• modifying the proposal;

• delaying the decision to gather information;

• using a less intrusive or less harmful alternative;

• sharing responsibility or decision authority;

• compensating or protecting those who bear unusual burdens.

Next, identify the values and duties at stake. Common values include autonomy, privacy, honesty, safety, welfare, equality, justice, loyalty, accountability, sustainability, and respect for human dignity. Common duties include keeping promises, preventing avoidable harm, treating people fairly, protecting confidentiality, and correcting wrongdoing.

A trade-off occurs when advancing one value limits another. Increased surveillance may improve security while reducing privacy. The existence of a trade-off does not prove that either value should be abandoned. Ask instead:

• Is the burden necessary rather than merely convenient?

• Is the response proportionate to the problem?

• Are burdens and benefits distributed fairly?

• Are affected people informed and able to challenge the decision?

• Are there safeguards, alternatives, or ways to reduce the harm?

• Who is accountable if the decision produces an unjust result?

Only after the issue has been framed should different ethical perspectives be applied. Consequentialist reasoning emphasizes effects on affected parties; deontological reasoning emphasizes duties and rights; virtue ethics asks what a good or practically wise person would do; and care ethics gives special attention to relationships, dependence, and vulnerability.

Takeaway: A defensible ethical analysis makes alternatives, values, conflicts, safeguards, and accountability visible before reaching a conclusion.

Applying the Method to Automated Hiring

Consider an employer proposing an automated system to rank job applications. The system could reduce processing time, but its training data may reflect historical hiring patterns.

Begin by identifying the facts that matter:

• What data were used to train the system?

• How accurate is it compared with human screening?

• Does performance differ across demographic groups?

• Can applicants request review or correction?

• Who is accountable when the system makes an error?

• Is the system an aid to human judgment or the final decision-maker?

Clarify the concepts involved:

• Does fairness require equal selection rates, equal opportunity, or attention to job-relevant qualifications?

• What makes a decision sufficiently explainable?

• What level of error is acceptable when employment opportunities are at stake?

Identify the affected groups:

• job applicants, including rejected applicants;

• current employees whose data may be used;

• hiring managers;

• the employer and its owners;

• groups historically disadvantaged in employment;

• customers and the wider community affected by workforce composition.

The issue can then be framed as follows:

Should the employer use the automated screening system, and under what conditions, when it may improve efficiency but may also reproduce discriminatory patterns and make decisions that applicants cannot understand or challenge?

This framing does not provide an automatic answer. It creates a basis for evaluating safeguards such as validation, human review, transparency, appeal procedures, and ongoing monitoring.

Takeaway: A worked frame should connect evidence, concepts, stakeholders, competing values, and realistic safeguards without presupposing the final decision.