5. Judgment, Decision Making and Behavioral Ethics

Introduction: Why Organizational Judgment Matters

Organizations are coordinated systems of choices. Leaders decide which opportunities to pursue, how resources should be allocated, whom to appoint, which risks to accept, and how competing stakeholder claims should be reconciled. Although organizational decisions are often represented as analytical exercises, they are also social, emotional, political, and ethical events. The quality of a decision therefore depends not only on the information available but also on how decision makers interpret that information, frame alternatives, manage uncertainty, and understand their responsibilities.

Judgment refers to the process of forming evaluations, estimates, or conclusions from available information. Decision making is the process of selecting and implementing a course of action. Behavioral ethics examines how individuals and groups actually behave when ethical values conflict with incentives, identity, authority, loyalty, or organizational pressure. These topics are closely connected. A flawed judgment can produce a poor decision; a poorly designed decision process can conceal ethical consequences; and an ethically compromised environment can distort both judgment and choice.

Classical decision theory assumes that decision makers identify objectives, compare alternatives, calculate consequences, and choose the option that maximizes expected value. This rational model remains valuable as a normative standard, but real organizational behavior rarely conforms to it completely. Managers operate with incomplete information, limited time, competing goals, cognitive constraints, and ambiguous feedback. They consequently rely on simplified representations, rules of thumb, experience, emotion, and social cues.

A doctoral understanding of decision making requires more than cataloguing biases. It requires examining the assumptions behind decision models, distinguishing descriptive from normative theories, identifying the conditions under which particular heuristics are useful or dangerous, and designing organizational systems that improve judgment without eliminating human responsibility.

Learning Objectives

By the end of this chapter, the reader should be able to:

1. Distinguish rational, bounded-rationality, intuitive, and behavioral models of decision making.

2. Explain how uncertainty, ambiguity, information quality, and framing affect organizational judgment.

3. Diagnose major cognitive and social biases, including anchoring, availability, confirmation, overconfidence, escalation of commitment, and groupthink.

4. Analyze ethical fading, moral disengagement, conflicts of interest, and obedience to authority.

5. Apply structured methods for improving decisions in leadership and organizational settings.

6. Evaluate the roles of data analytics, artificial intelligence, governance, and human oversight in contemporary decision processes.

7. Critically assess decision failures and derive lessons for organizational design and professional practice.

The Architecture of Organizational Choice

A decision begins with a problem representation. The same situation may be described as a cost problem, a quality problem, a competitive threat, or an ethical issue. The framing selected by a leader determines which information appears relevant and which alternatives are considered legitimate. Problem identification is therefore not a neutral preliminary step; it is itself a judgment.

Decision environments can be classified according to the relationship between alternatives, consequences, and probabilities. Under certainty, consequences are assumed to be known. Under risk, several outcomes are possible and probabilities can be estimated. Under uncertainty, probabilities are unavailable, disputed, or unreliable. Ambiguity is more profound: decision makers may disagree about what the possible outcomes even are or how they should be interpreted. Most strategic and leadership decisions involve a mixture of risk, uncertainty, and ambiguity.

The quality of a decision should also be distinguished from the quality of its outcome. A sound decision may produce an unfavorable result because of chance, while a careless decision may succeed accidentally. Evaluating decisions solely by outcomes encourages hindsight bias and can cause organizations to reward reckless luck. Process quality, information available at the time, assumptions made, and the reasonableness of the analysis must also be examined.

The Rational Model

The rational model is a normative account of how a fully informed and logically consistent decision maker should act. Its typical sequence is:

• define objectives and the problem;
• identify alternatives;
• gather relevant information;
• estimate consequences and probabilities;
• compare alternatives using explicit criteria;
• select the option with the greatest expected utility; and
• implement and evaluate the decision.

Expected utility is often expressed as the sum of each possible outcome multiplied by its probability and value. This model supports cost-benefit analysis, scenario planning, investment appraisal, and formal risk management. It makes assumptions visible and allows alternatives to be compared systematically.

Its limitations are equally important. Objectives may be contradictory or difficult to quantify. Relevant information may be unavailable or expensive. Probabilities may be subjective. Decision makers may manipulate criteria to justify a preferred option, and implementation may alter the conditions on which the original analysis was based. Rational analysis is therefore best treated as a discipline for improving judgment, not as proof that a choice is objectively optimal.

Bounded Rationality and Satisficing

Herbert Simon’s concept of bounded rationality explains why actual decision makers cannot achieve the ideal of comprehensive optimization. Human cognition, time, attention, and computational capacity are limited. Organizations also impose constraints through hierarchy, routines, budgets, and political interests. Rather than evaluating every possible alternative, managers construct simplified models of reality and search until they find an option that meets an acceptable threshold. Simon called this process satisficing: selecting an alternative that is satisfactory and sufficient, even if it is not demonstrably optimal.

Bounded rationality does not imply irrationality. Simplification is often necessary. A hospital administrator cannot investigate every possible supplier, and a leader responding to a crisis cannot conduct a full strategic review. The managerial challenge is to know which assumptions may safely be simplified and which deserve deeper scrutiny. Decision protocols, independent review, and staged commitments can expand the effective boundaries of rationality.

Intuition and Expertise

Intuitive judgment is a rapid, largely nonconscious assessment that draws on patterns learned through experience. It is not necessarily arbitrary or mysterious. In stable environments with regular feedback, experienced professionals can recognize meaningful configurations quickly. A seasoned engineer may detect an unusual system sound, or an emergency physician may recognize a dangerous clinical pattern before all data are available.

Intuition becomes less reliable when environments are volatile, feedback is delayed or distorted, patterns are rare, or incentives encourage wishful interpretation. Executive intuition can be especially vulnerable to status, ego, and selective experience. The appropriate question is not whether intuition should be used, but whether the setting supports valid expertise. A useful test asks whether the environment has stable regularities and whether the decision maker has received repeated, accurate feedback about previous judgments.

Behavioral Decision Theory

Behavioral decision theory combines insights from psychology and economics to describe how people actually choose. Prospect theory is particularly influential. It proposes that people evaluate outcomes relative to a reference point rather than absolute wealth, experience losses more intensely than equivalent gains, and often distort probabilities. This helps explain why leaders may reject a certain modest gain while accepting a risky gamble to avoid realizing a loss.

Framing is central. A restructuring described as saving 80 percent of jobs may receive greater support than one described as losing 20 percent, even though the statements are logically equivalent. Behavioral decision theory therefore directs attention to the language, comparisons, timing, and emotional meaning through which alternatives are presented.

Heuristics, Biases, and Social Distortion

Heuristics are cognitive shortcuts that reduce the effort required to make judgments. They can be efficient and adaptive, but they can also produce systematic error. A bias is not simply any error; it is a predictable tendency that influences judgment in a particular direction.

Anchoring occurs when an initial number, proposal, or interpretation exerts excessive influence on subsequent judgment. An opening price can shape negotiations even when it is arbitrary. Leaders can reduce anchoring by generating independent estimates before group discussion and by asking what evidence would justify a substantially different estimate.

Availability bias arises when vivid, recent, or emotionally striking events are judged as more probable than they really are. After a highly publicized cyberattack, an organization may overinvest in that threat while neglecting less visible vulnerabilities. Incident registers and base-rate data help restore proportion.

Representativeness bias involves judging probability by similarity to a familiar category while neglecting statistical base rates. A confident candidate may be perceived as a natural leader despite weak evidence of relevant performance. Structured criteria and validated selection procedures counter this tendency.

Confirmation bias is the inclination to seek, interpret, and remember evidence that supports an existing belief. It is particularly dangerous in strategy, where senior leaders may discourage information that challenges a favored vision. A decision process should assign someone to develop the strongest case against the leading proposal.

Overconfidence includes excessive certainty in estimates, predictions, or personal ability. It can produce unrealistic schedules, insufficient reserves, and excessive acquisition prices. Calibration exercises, reference-class forecasting, and explicit confidence ranges can expose unjustified certainty.

Action bias is the tendency to prefer doing something over waiting, even when inaction or information gathering would be wiser. Conversely, omission bias may lead managers to avoid a visible action because its consequences would be personally attributable. Both tendencies reflect accountability pressures rather than objective analysis.

Status quo bias favors existing arrangements, while loss aversion makes potential losses loom larger than equivalent gains. Together they can preserve obsolete systems and delay necessary transformation.

Groupthink occurs when the desire for cohesion suppresses dissent, critical evaluation, and consideration of alternatives. Symptoms include an illusion of unanimity, stereotyping of outsiders, self-censorship, and pressure on dissenters. Groupthink is not simply agreement; agreement reached after rigorous challenge may be a sign of high-quality deliberation.

Escalation of commitment occurs when individuals or groups continue investing in a failing course of action because of sunk costs, identity, reputation, or the desire to justify earlier choices. A project may be defended because abandoning it would imply that senior leaders were wrong. Decision gates, predetermined exit criteria, independent reviews, and separation of project advocacy from continuation decisions reduce escalation.

Biases rarely operate alone. An anchor can influence the information searched; confirmation bias can reinforce an initial interpretation; group loyalty can silence objections; and escalation can make later evidence appear threatening. Diagnosis must therefore examine the whole decision environment rather than identify one isolated cognitive error.

Ethical Judgment and Ethical Fading

Behavioral ethics asks why people who regard themselves as moral sometimes make choices that harm others or violate professional standards. Ethical behavior is not produced solely by personal character. It is shaped by incentives, role expectations, authority relationships, organizational culture, time pressure, and the way problems are framed.

Ethical fading occurs when the moral dimensions of a decision disappear from attention. A sales target may be described only as a performance metric, a customer-data practice as an efficiency initiative, or a discriminatory algorithm as a technical optimization problem. Once ethical language fades, individuals can act contrary to their values without experiencing the decision as an ethical choice.

Moral disengagement provides mechanisms through which harmful conduct is justified. Individuals may use euphemistic language, diffuse responsibility across a team, minimize consequences, blame victims, or compare their conduct with something worse. Conflicts of interest can also distort judgment, especially when the conflict is disclosed but not effectively managed. Disclosure alone does not remove bias; recusal, independent review, or structural separation may be required.

Authority and obedience create further risks. Employees may comply with questionable instructions because responsibility appears to rest with a superior. Leaders therefore have a special obligation to make ethical concerns discussable, protect dissent, and clarify that obedience does not excuse violations of law, professional standards, or human dignity.

An ethical decision process can be organized around several questions:

1. Who may benefit, bear costs, or be exposed to risk?

2. What rights, duties, professional standards, or laws are relevant?

3. Is the action fair in its process and distribution of benefits and burdens?

4. Would the decision remain defensible if publicly disclosed?

5. What would an impartial observer, affected stakeholder, or future employee say about it?

6. Can the decision be explained without euphemism or concealment?

These questions draw on consequentialist, deontological, justice-based, virtue-ethical, and stakeholder perspectives. No single test resolves every dilemma, but the combination makes ethical implications harder to ignore.

Improving Decision Quality in Organizations

Good decision architecture does not attempt to eliminate judgment. It places judgment where it is most useful and creates safeguards against predictable failure. The following practices are particularly valuable.

Separate stages of analysis. Problem definition, option generation, evaluation, and authorization should not collapse into one conversation. Separating them reduces premature commitment.

Use independent estimates. Before discussion, participants should record forecasts, assumptions, and confidence levels. This prevents the first speaker or most senior person from anchoring the group.

Conduct a premortem. The team imagines that the decision has failed and identifies plausible reasons. This legitimizes criticism before implementation and reveals hidden assumptions.

Seek disconfirming evidence. Assign a devil’s advocate or red team to develop the strongest credible challenge. The role should rotate so that dissent is treated as a process responsibility rather than personal opposition.

Use reference classes. Compare the proposal with outcomes of similar projects rather than relying exclusively on an internal narrative. This counters the planning fallacy, in which managers underestimate time, cost, and difficulty.

Define decision rights and stop rules. Before committing resources, specify who decides, what evidence triggers review, and when a project should be paused or terminated. Precommitment is particularly important where escalation is likely.

Create psychological safety. Employees must be able to raise concerns without humiliation or retaliation. Psychological safety does not mean avoiding accountability; it enables earlier disclosure of error and dissent.

Document assumptions and conduct after-action reviews. Records should distinguish what was known at the time from information discovered later. Reviews should examine both decision process and implementation, avoiding blame-oriented hindsight.

Match analysis to the decision. High-stakes, irreversible, and ethically sensitive decisions deserve more scrutiny than routine, reversible choices. Excessive analysis can also delay action, so decision processes should be proportionate to consequence and uncertainty.

Data, Algorithms, and Human Oversight

Digital tools can improve judgment by detecting patterns, aggregating evidence, modeling scenarios, and reducing inconsistent application of rules. Predictive analytics may support workforce planning, fraud detection, supply-chain management, or credit assessment. Yet data-driven decisions do not automatically become objective. Historical data may encode discrimination, variables may act as proxies for protected characteristics, and optimization objectives may reward what is measurable rather than what is valuable.

Automation bias occurs when users over-rely on an algorithmic recommendation, even when contextual evidence indicates that it is wrong. The opposite danger is algorithm aversion: rejecting useful tools because of a visible error while ignoring comparable human errors. Effective governance requires clear accountability, data-quality controls, explainability appropriate to the context, monitoring for disparate impacts, privacy protection, and meaningful human review.

Generative and predictive artificial intelligence intensify these issues. Leaders must distinguish prediction from judgment, correlation from causation, and statistical performance from ethical legitimacy. Human oversight should not be merely ceremonial. Reviewers need authority, competence, time, and access to the information required to challenge a system.

Illustrative Case: The Meridian Automation Decision

Background

Meridian Components, a fictional manufacturer, considered replacing a manual quality-inspection process with an artificial-intelligence vision system. The executive sponsor predicted a 25 percent reduction in inspection costs and presented the project as essential to remaining competitive. The board approved a pilot after a short presentation based largely on vendor demonstrations.

Situation and Problem

During the pilot, the system performed well on standard products but missed defects in a newly introduced product line. Production managers were concerned, but the vendor argued that additional training data would solve the problem. The chief operating officer had already announced publicly that automation would deliver major savings. Engineers therefore hesitated to recommend pausing the project. Meanwhile, the compliance officer raised concerns that quality failures could affect customer safety and that the training data did not represent all production conditions.

Analysis

Several behavioral and ethical problems were present. The initial cost-saving estimate served as an anchor. Confirmation bias led managers to emphasize successful test results and discount failures. Escalation of commitment developed because money and executive reputation had already been invested. Ethical fading occurred when the discussion was framed as an efficiency project rather than a decision affecting product safety. Authority pressure discouraged engineers from challenging the sponsor. The vendor’s performance claims also lacked an independent reference class and confidence range.

Approach and Solution

Meridian established an independent review team consisting of quality engineers, production staff, data specialists, legal advisers, and a customer representative. The team separated the questions of technical accuracy, financial value, and ethical acceptability. It conducted a premortem, tested the system on disaggregated product and operating conditions, and established a stop rule: no full deployment would occur until error rates met predefined safety thresholds across all relevant categories.

The board also required dual control during the transition. Human inspectors remained responsible for final release, and all disagreements between the system and inspectors were logged. The vendor contract was revised to require audit access, data documentation, and notification of model changes. After further testing, Meridian adopted the system for low-risk products but retained manual inspection for the new line until more representative data became available.

Outcome

The organization achieved smaller initial savings than forecast, but it avoided a premature full-scale deployment and identified a defect pattern that the original pilot had concealed. The decision was recorded as a staged adoption rather than a rejection of innovation. Later reviews assessed performance, safety, workforce effects, and customer complaints separately.

Learning Points

The case demonstrates that decision quality improves when advocates do not control evaluation, assumptions are made explicit, and ethical consequences are treated as part of the decision rather than as a compliance afterthought. It also shows why human oversight must include authority to stop implementation, not merely responsibility to observe it.

Limits and Common Failure Modes

Decision frameworks can themselves become rituals. A checklist may be completed mechanically while senior leaders have already decided the outcome. A devil’s advocate may be ignored if dissent is socially punished. Data can create false precision, and formal models may conceal value judgments behind technical language. Ethical codes may be ineffective if reward systems favor results regardless of methods.

Another danger is bias overdiagnosis. Not every unfavorable outcome demonstrates bias, and not every disagreement reflects groupthink. Researchers and practitioners should use evidence, compare alternative explanations, and avoid treating psychological labels as substitutes for analysis. The language of bias can also become a way of blaming individuals for failures created by organizational incentives.

Leaders should be alert to the interaction between culture and structure. A culture that celebrates certainty may suppress learning; one that rewards heroic crisis management may create action bias; one that treats missed targets as personal failure may encourage concealment. Effective governance therefore combines individual reflection with institutional redesign.

Current Developments and Research Directions

Contemporary research increasingly examines judgment as a property of teams, networks, and institutions rather than isolated individuals. Attention is shifting toward collective intelligence, distributed expertise, dissent quality, and the design of information flows. Behavioral insights are also being applied to climate risk, public policy, cybersecurity, and responsible innovation, where low-probability events and delayed consequences complicate conventional analysis.

A growing research concern is the governance of algorithmic decision systems. Questions include who is accountable when responsibility is distributed across developers and users, how fairness should be defined across groups, and how organizations can audit systems that change over time. Another emerging direction is the study of decision resilience: the capacity of organizations to detect weak signals, revise assumptions, learn from near misses, and recover without defensiveness.

Best-Practice Principles for Leaders

Leaders who seek more reliable and ethical decisions should:

• make the problem and its moral stakes explicit;
• distinguish facts, assumptions, forecasts, and values;
• invite independent estimates before discussion;
• reward informed dissent and protect psychological safety;
• use base rates, reference classes, and confidence intervals;
• predefine review points, exit criteria, and decision rights;
• separate advocacy from evaluation;
• assess distributional effects, rights, and stakeholder consequences;
• treat algorithms as aids to judgment, not substitutes for accountability; and
• review decisions in a way that learns without relying on hindsight blame.

Chapter Summary

Judgment and decision making are central leadership activities shaped by cognition, emotion, social influence, organizational structure, and ethics. The rational model provides a useful normative benchmark, but bounded rationality explains why managers simplify, satisfice, and rely on routines. Intuition can be effective when experience is genuine and feedback is reliable, yet it is hazardous in unstable or ambiguous environments. Behavioral decision theory demonstrates how framing, reference points, loss aversion, and probability distortion influence choice.

Heuristics can make decisions efficient, but anchoring, availability, representativeness, confirmation, overconfidence, action bias, groupthink, and escalation of commitment can generate systematic failure. Ethical fading and moral disengagement show how organizational pressures can conceal moral issues from otherwise conscientious people. Decision quality can be improved through independent estimates, premortems, red teams, reference classes, stop rules, psychological safety, transparent documentation, and proportionate review. Data and artificial intelligence offer valuable support but require attention to bias, explainability, privacy, automation bias, and human accountability. Ultimately, responsible leadership requires designing decision environments in which rigorous analysis, informed intuition, dissent, and ethical reflection reinforce one another.

Key Terms

Anchoring: Excessive influence of an initial number, idea, or interpretation on later judgment.

Availability bias: Judging likelihood by the ease with which examples come to mind.

Bounded rationality: Decision making constrained by limited information, time, cognition, and organizational capacity.

Confirmation bias: Seeking or interpreting evidence in ways that support existing beliefs.

Decision quality: The soundness of the process and reasoning used, considered separately from the eventual outcome.

Ethical fading: The disappearance of moral considerations from the way a decision is framed.

Escalation of commitment: Continued investment in a failing course because of sunk costs, identity, or reputation.

Expected utility: The probability-weighted value of possible outcomes.

Framing effect: Different choices resulting from how equivalent information or alternatives are presented.

Groupthink: Suppression of dissent and critical evaluation in pursuit of group cohesion.

Heuristic: A simplifying rule or mental shortcut used for judgment and choice.

Intuition: Rapid judgment based largely on learned patterns and experience.

Loss aversion: The tendency for losses to have greater psychological impact than equivalent gains.

Moral disengagement: Cognitive processes that make harmful conduct appear acceptable or responsibility-free.

Overconfidence: Excessive certainty in one’s judgments, predictions, or capabilities.

Premortem: An exercise in which a team imagines failure and identifies its plausible causes before action.

Prospect theory: A behavioral theory proposing reference-dependent evaluation, loss aversion, and distorted probability judgment.

Satisficing: Selecting an option that meets an acceptable threshold rather than optimizing across all possibilities.

Uncertainty: A condition in which outcomes or their probabilities are unknown, disputed, or unreliable.

Review Questions

Short Answer Questions

1. How does bounded rationality differ from the rational decision model?

2. Under what conditions is managerial intuition most likely to be reliable?

3. What is ethical fading, and how can organizational language produce it?

4. Distinguish escalation of commitment from ordinary persistence.

5. Why should decision quality be evaluated separately from decision outcomes?

Long Answer Questions

1. Critically compare rational, bounded-rationality, intuitive, and behavioral decision models.

2. Explain how anchoring, confirmation bias, and overconfidence can interact in strategic decision making.

3. Discuss the relationship between organizational culture, psychological safety, and ethical judgment.

4. Evaluate the usefulness and limitations of premortems, red teams, reference classes, and stop rules.

5. Analyze the governance challenges created when artificial intelligence supports high-stakes organizational decisions.

Analytical / Application Questions

1. A firm continues funding a product that has missed every major milestone, largely because its CEO publicly promised success. Diagnose the decision failures and design an intervention.

2. A hiring algorithm selects fewer candidates from a historically underrepresented group, although its overall accuracy is high. How should leaders investigate and govern its use?

3. A crisis team must decide whether to delay a product launch despite incomplete evidence of a safety problem. Apply at least two ethical frameworks and a structured decision process to the situation.

22. Multiple Choice Questions with Answers

Question 1

Which statement best captures bounded rationality?

A. Managers always maximize expected utility when sufficient incentives exist.

B. Managers make choices under limits of information, time, cognition, and organizational capacity.

C. Managers rely exclusively on emotional reactions rather than analysis.

D. Managers avoid making decisions unless all probabilities are known.

Correct Answer: B. Managers make choices under limits of information, time, cognition, and organizational capacity.

Question 2

A leader estimates that a project will take six months because similar projects in the organization usually take six months, despite evidence that comparable industry projects take much longer. Which corrective method is most appropriate?

A. Increase commitment to the current estimate

B. Use reference-class forecasting and external base rates

C. Ask the project sponsor to express greater confidence

D. Remove all uncertainty from the project plan

Correct Answer: B. Use reference-class forecasting and external base rates

Question 3

Which situation most clearly illustrates ethical fading?

A. A team debates whether a decision is consistent with stakeholder rights.

B. A manager consults an ethics officer before changing supplier standards.

C. A customer-data initiative is discussed only as an efficiency project, with privacy consequences ignored.

D. An employee reports a conflict of interest to an independent committee.

Correct Answer: C. A customer-data initiative is discussed only as an efficiency project, with privacy consequences ignored.

Question 4

Why can escalation of commitment continue after evidence indicates that a project is failing?

A. Past investments and public commitments make withdrawal psychologically and politically costly.

B. Decision makers always receive better information after investing more resources.

C. Sunk costs automatically increase the project’s expected value.

D. Independent review eliminates the influence of identity and reputation.

Correct Answer: A. Past investments and public commitments make withdrawal psychologically and politically costly.

Question 5

Which practice most directly reduces anchoring during a group decision?

A. Allowing the most senior participant to speak first

B. Requiring participants to record independent estimates before discussion

C. Using only the first available forecast

D. Avoiding quantitative information altogether

Correct Answer: B. Requiring participants to record independent estimates before discussion

Question 6

A decision maker rejects a certain moderate gain but accepts a risky option because it offers a chance to recover an earlier loss. This pattern is most consistent with:

A. Loss aversion and prospect-theory framing

B. The elimination of bounded rationality

C. Purely objective expected-utility maximization

D. Availability bias alone

Correct Answer: A. Loss aversion and prospect-theory framing

Question 7

Which feature distinguishes productive dissent from groupthink?

A. Dissent is suppressed until consensus is reached.

B. Agreement is accepted only after alternatives and counterarguments have been examined.

C. The leader announces the preferred option before analysis begins.

D. Team members avoid documenting disagreements.

Correct Answer: B. Agreement is accepted only after alternatives and counterarguments have been examined.

Question 8

Why is human oversight of an algorithmic decision system potentially inadequate if reviewers lack authority to intervene?

A. Oversight then becomes observational rather than accountable and corrective.

B. Human reviewers will necessarily reject every algorithmic recommendation.

C. Algorithms cannot process any organizational data.

D. Authority always increases automation bias.

Correct Answer: A. Oversight then becomes observational rather than accountable and corrective.

Question 9

Which action best addresses a conflict of interest that may distort judgment?

A. Relying only on voluntary disclosure

B. Increasing the individual’s personal stake in the outcome

C. Using recusal, independent review, or structural separation where appropriate

D. Keeping the conflict confidential to preserve team harmony

Correct Answer: C. Using recusal, independent review, or structural separation where appropriate

Question 10

In the Meridian case, why was staged adoption preferable to immediate full deployment?

A. It eliminated the need for human responsibility.

B. It allowed evidence, safety thresholds, and review points to guide commitment under uncertainty.

C. It guaranteed the vendor’s original savings estimate.

D. It prevented any future use of artificial intelligence.

Correct Answer: B. It allowed evidence, safety thresholds, and review points to guide commitment under uncertainty.