Behavioral Ethics: S-Tier Behavioral Designer’s Guide

In 2003, two researchers at Harvard and Notre Dame did something the field of ethics had largely refused to do for two thousand years. They stopped asking what people should do and started measuring what people actually do, including the people who, asked in advance, swore they would do the right thing. The gap they uncovered is wider than almost anyone in business or policy is comfortable acknowledging, and it explains why every “ethics training” deck since 2003 has produced the same disappointing results.

The gap has a name. Max Bazerman and Ann Tenbrunsel call it bounded ethicality, and they built it as the deliberate analogue to Herbert Simon’s bounded rationality. Just as Simon showed that our reasoning is constrained by cognitive limits we can’t see, Bazerman and Tenbrunsel showed that our ethics are constrained by psychological forces we can’t see. Those forces systematically pull good people into decisions they would have rejected if they’d been asked beforehand. The forces aren’t character flaws. They’re predictable, repeatable cognitive mechanisms. And once you can name them, you can design against them.

If you build commercial products, run a team, sit on a board, or own a personal brand, this is the lens you’ve been missing. Behavioral ethics is what happens when Octalysis meets Grey Hat: every Core Drive that motivates engagement is also, in some form, a vector for systematic ethical failure. The same psychological machinery that makes a loyalty program magnetic is the machinery that built the Wells Fargo cross-sell scandal. The trick is learning to see both at once.

⚡ Speed Run Notes

• Behavioral ethics is the descriptive study of how good people end up doing bad things. Not bad people; good people, under predictable psychological pressures they cannot see.
• Bounded ethicality means our ethical reasoning is constrained the way our cognitive reasoning is. Five mechanisms drive the gap: ethical fading, motivated blindness, indirect blindness, slippery slope, and outcome bias.
• The Want Self beats the Should Self the moment money, time pressure, or status are on the table. Predictions and recollections agree we will act ethically; only the action itself betrays us.
• Every Octalysis Core Drive that motivates engagement is also a bounded-ethicality vector. Aggressive goal-setting builds Wells Fargo. Time pressure builds Boeing. Loss avoidance builds the cover-up.
• Ethics training that targets the Should Self misses the problem. The fix is environmental: change the choice architecture so the Want Self never gets to vote in private.
• The Octalysis-Grey-Hat audit names where each Core Drive becomes ethically dangerous, then redesigns the moment of decision rather than the morality of the decider.

Table of Contents

About Yu-kai Chou

Yu-kai Chou created the Octalysis Framework after studying gamification since 2003 — years before the term entered mainstream vocabulary. As a Human-Systems Architect & Behavioral Designer, his framework has been applied by LEGO, Microsoft, Porsche, Coca-Cola, Salesforce, and MrBeast, impacting over 1.5 Billion Users.

Chou has taught the Octalysis methodology at Harvard, Stanford, Yale, Tesla, Google, BCG, and IDEO.

His work has been cited by Harvard, Stanford, MIT, Forbes, Wall Street Journal, Wired, US Department of Energy, NIST, NSF, NCBI, US Department of Education, ClinicalTrials.gov, and Google Scholar — with 3,700+ more academic publications. Explore his books here.

TEDxLausanne · Gamification to Improve Our World17 min

What Is Behavioral Ethics?

Classical ethics, from Kant through Mill to Rawls, asks one question: what is the right thing to do? The discipline has spent two millennia building elegant frameworks to answer it. The trouble with elegant frameworks is that they describe a creature humans do not actually resemble. They describe a deliberative agent who pauses, consults principles, weighs consequences, and chooses. Behavioral ethics describes the agent we actually are: pressed for time, embedded in a social context, financially exposed, and quietly motivated to see things in a way that serves us.

Behavioral ethics is the descriptive study of how good people end up making unethical decisions they would have rejected if asked beforehand. Bazerman and Tenbrunsel published the definitive book on it, Blind Spots: Why We Fail to Do What’s Right and What to Do about It, in 2011. The field draws on three decades of laboratory and field research by Bazerman, Tenbrunsel, David Messick, Dolly Chugh, Mahzarin Banaji, Francesca Gino, Dan Ariely, and others, and it sits at the intersection of social psychology, behavioral economics, and organizational behavior.

The central claim is small and unsettling. Most unethical behavior in organizations is not committed by people who decided to be unethical. It is committed by people who, in the moment of decision, did not recognize the decision as having an ethical dimension at all. They saw a business problem, a deadline, a quota, a defensible legal interpretation, a customary practice. The ethical content of the decision faded from view long before the choice was made, and it returned only afterward, when the decider, often genuinely confused, tried to explain what happened.

This positions behavioral ethics in deliberate opposition to two older programs. The first is normative ethics, which prescribes what we should do; behavioral ethics is descriptive, mapping what we actually do. The second is the bad-apple theory of organizational scandals, which locates failure in the moral character of a few rogue individuals; behavioral ethics says the problem is the barrel, not the apples, and that the barrel is built out of psychological mechanisms operating below conscious awareness.

The implication for design is significant. If unethical behavior is mostly a function of context rather than character, then better behavior is mostly an architectural problem. You don’t fix it with more ethics training. You fix it by changing the moment of decision so the predictable psychological shortcuts don’t get a chance to fire.

The Five Mechanisms of Bounded Ethicality

Bounded ethicality is the umbrella term. The machinery underneath consists of five distinct mechanisms, each well documented, each capable of producing massive failures on its own, and each amplified by the others when they combine in the wild.

Ethical Fading

Ethical fading is the psychological process where the ethical dimension of a decision recedes from awareness, replaced by frames that do not register as ethical. Tenbrunsel and David Messick documented the mechanism in a 2004 paper that has since become the field’s most-cited definition. The frame that does the replacing is usually a business frame, a legal frame, or a competitive frame. Once the new frame is in place, the decider is not weighing right against wrong; they are weighing efficiency, exposure, or strategy.

The clearest experimental demonstration comes from a study Tenbrunsel ran where MBA students were given a CFO decision about whether to use an accounting practice that was legal but ethically questionable. When the choice was framed as an ethical decision, students chose conservatively. When the same choice was framed as a business decision, the same students chose aggressively. The decision had not changed. The frame around it had. The frame did the work.

In the wild, ethical fading is what produces the engineer at Volkswagen who calls the emissions cheat device an “acoustic function,” the financial analyst who calls toxic mortgage tranches “structured products,” and the casino executive who calls problem gamblers “high-value VIPs.” The substitute language is not deception aimed at outsiders; it is the language the insider uses with themselves, and it is what allows the insider to act without registering the act as something ethical.

Motivated Blindness

Motivated blindness is the systematic failure to notice information that, if noticed, would oblige us to do something inconvenient. Bazerman and Don Moore documented the mechanism most famously in the auditor independence literature. Auditors paid by the firm they audit do not consciously falsify their reports. They simply, reliably, fail to see things they would have seen if they were paid by an independent party. The bias is not corrupt judgment; it is corrupt attention.

The most influential demonstration of the underlying mechanism was the work of Don Moore, George Loewenstein, Lloyd Tanlu, and Bazerman, who showed in a 2006 paper that conflict of interest produces a systematic blindspot that survives both training and incentives for accuracy. Auditors told their bias did not improve their accuracy. Auditors with stronger ethical training did not improve their accuracy. The only variable that mattered was the conflict itself.

The applied version: every gamification system that pays out on engagement metrics is paying its designers to be motivatedly blind to the addiction tail of those metrics. Every analytics dashboard that promotes north-star numbers makes the side effects of the north-star numbers invisible. Every consultant paid on a per-engagement basis is, by structure, motivatedly blind to the question of whether the engagement was the right kind.

Indirect Blindness

Indirect blindness is the tendency to judge unethical conduct less harshly when the conduct happens through intermediaries. The mechanism was formalized in a 2011 paper by Lisa Shu, Francesca Gino, and Bazerman. Subjects asked to judge a CEO who raises a drug price 600 percent rate the conduct as far worse when the CEO does it directly than when the CEO sells the drug to a smaller company that then raises the price. The conduct is identical; the responsibility chain has one extra link.

This is the mechanism behind a great deal of corporate misconduct that nobody at the principal company will admit to authorizing. Garment factories in Bangladesh, content moderators in the Philippines, ride-share driver acquisition through aggressive lead-gen firms, debt collectors retained by banks — the intermediary structure is, among other things, a moral-distance amplifier. The principal does not see the conduct. The intermediary does not see the principal’s intent. Each party, individually, can describe their own conduct as defensible. The aggregate is what produces the harm.

Slippery Slope (Incrementalism)

The slippery slope is the tendency for small ethical violations to compound into massive ones because each step is small enough to be invisible. Gino and Bazerman documented the mechanism in a 2009 paper showing that auditors who would have rejected a single large overstatement readily accepted a series of small overstatements that summed to the same large number. The total was identical. The path was different. The path did the work.

The mechanism has two pieces. First, each small step does not register as a meaningful ethical violation. Second, having taken the small step, the decider is now invested in not seeing it as a violation, because seeing it would require revising their self-image as an ethical person. The two pieces interact: invisibility breeds investment, investment breeds further invisibility. By the time the cumulative misconduct is large enough to be visible from outside, the decider is too committed to their innocence to recognize it from inside.

Enron’s accounting did not become Enron’s accounting in one decision. It became Enron’s accounting through a thousand decisions, each of which felt, in the moment, like a continuation of the previous one. The Wells Fargo cross-sell scandal followed the same arc: each individual unauthorized account felt like a defensible work-around for a quota that was, itself, the result of a defensible business strategy. The slope is built one defensible step at a time.

Outcome Bias

Outcome bias is the tendency to judge the ethics of a decision based on the outcome rather than the process. Baron and Hershey first documented it in 1988; Bazerman, Tenbrunsel, and others extended it specifically to ethical judgments. The same risk taken with the same information is judged as more ethical when the outcome is good than when the outcome is bad. The decision did not change. The result did. The result did the moral work.

The applied consequence is severe. Outcome bias means that organizations that escape the worst consequences of their behavior conclude, retrospectively, that the behavior was acceptable. It means that retrospective ethics reviews systematically punish bad luck and reward good luck. It means that organizational learning, in the ethics domain, is biased toward the wrong lessons. The most dangerous executives are the ones whose previous high-risk gambles paid off, because their own track record tells them — falsely — that their judgment is sound.

The Want Self vs. Should Self Architecture

Behind the five mechanisms sits a deeper structural claim, and it is the part of behavioral ethics that translates most directly into design. Bazerman and Tenbrunsel argue that ethical decisions move through three phases — prediction, action, recollection — and that a different version of the self runs each phase. The Should Self runs the prediction phase. The Want Self runs the action phase. The Should Self returns for the recollection phase and rewrites the action to fit its self-image. The three phases are inhabited by what look like three different people, and the action-phase person is the one who actually controls the behavior.

Ask employees in advance whether they would falsify expense reports for a small personal benefit and the overwhelming majority say no. This is the Should Self speaking. Watch the same employees in the moment, under time pressure, in a context where the falsification is a minor variation on standard practice and the Want Self speaks. The two answers are not lies told by the same person at different times. They are accurate reports from two different cognitive systems — one deliberative, one reactive — that share a body but not a perspective.

The implication for organizational design is that interventions targeting the Should Self do not transfer to the Want Self. Every ethics training that asks employees, in a quiet room, what they would do in an unethical situation is talking to the Should Self. The training does not reach the Want Self because the Want Self is not in the room. The Want Self is at the desk, on the deadline, six hours into a difficult day, watching a colleague execute a defensible workaround. The intervention has to be present at that moment or it has no effect.

The asymmetry across the three phases is itself part of the trap. People predict ethical behavior accurately for others and unrealistically for themselves. They recall their own behavior more ethically than it actually was. The pattern is well-attested: we forecast the Should Self for our future, observe the Want Self in our present, and remember a smoothed version of the Want Self that looks more like the Should Self. The smoothing protects self-image but it also blocks learning, because if we cannot accurately recall the gap between intention and behavior, we cannot revise the next prediction.

The fix is environmental. If the Want Self is the actor and the Want Self is structurally short-sighted under pressure, then the design move is to make the ethical path the path of least resistance at the moment of action. Not to lecture the Should Self about a future Want Self it will not remember being. Not to remind the Want Self of principles the Should Self believes in. Change the moment.

What Bazerman & Tenbrunsel Got Right

The single biggest contribution of behavioral ethics is the relocation of the ethics problem from character to context. For two thousand years, the question of how to make people more ethical has been answered with various combinations of education, exhortation, and threat. Behavioral ethics says the answer to that question is the wrong answer to the wrong question. The right question is how to make the moment of decision survive the Want Self.

The second contribution is the descriptive map of the mechanisms. Before Bazerman and Tenbrunsel, the various failure modes of ethical reasoning lived in separate literatures — cognitive dissonance over here, motivated reasoning over there, framing effects in a third place. Behavioral ethics integrates them into a single architecture organized around when, where, and why the ethical content of a decision drops out of awareness. The integration is what makes the field practical. A designer can audit a system for ethical fading or motivated blindness in a way that they could not audit a system for “low ethics” in general.

The third contribution is the experimental rigor. Behavioral ethics is one of the few corners of applied ethics where the central claims are anchored in replicated laboratory and field studies, not in philosophical thought experiments. The Tenbrunsel-Messick ethical fading work, the Moore-Loewenstein-Tanlu-Bazerman auditor work, the Shu-Gino-Bazerman indirect blindness work, the Gino-Bazerman slippery slope work — each rests on multiple studies with adequately powered samples, mostly run before the replication crisis sent waves through social psychology, and mostly holding up in subsequent meta-analyses.

The fourth contribution is normative humility. Bazerman and Tenbrunsel are explicit that behavioral ethics does not displace normative ethics; it sits underneath it. You still need a theory of right and wrong. The descriptive field tells you how to design around the predictable ways that humans will fail to enact whatever normative theory you adopt. The two are complements, not substitutes. The book closes with explicit pointers back to Aristotle, Kant, and Mill for readers who want a theory of what to do; behavioral ethics is offered as a theory of how to do it.

Where Behavioral Ethics Falls Apart

The field is not above critique. Three pressure points are worth naming directly, because every commercial application of behavioral ethics has to grapple with.

The Diagnosis-Without-Cure Problem

Behavioral ethics is much stronger at diagnosing failure than at producing reliable interventions. The five mechanisms are well documented; the corresponding interventions are not. Reducing motivated blindness in auditors is conceptually simple — pay them independently — but structurally hard, because the entire audit industry runs on the wrong incentive. Reducing ethical fading requires reframing in-the-moment language at organizational scale, which fights every economic pressure toward jargon, euphemism, and legal-defensibility framing. The field has produced more useful warnings than useful fixes, and operators reading Blind Spots for actionable advice often find the actionable advice section thinner than the diagnosis section.

The Self-Application Paradox

If our ethical reasoning is bounded by mechanisms we cannot see, the people designing behavioral-ethics interventions are subject to the same bounded ethicality. There is a real risk that ethics consultants, ethics officers, and behavioral-ethics academics develop a professional motivated blindness toward the limits of their own field — toward the gap between what their interventions promise and what they deliver. The literature is starting to acknowledge this. Tenbrunsel herself has written about the limited efficacy of corporate ethics programs in a 2020 review. The honest position is that the field is still developing the discipline of auditing its own interventions for the same failures it warns about.

The Cultural Universality Question

Most foundational studies in behavioral ethics were run on WEIRD samples — Western, Educated, Industrialized, Rich, Democratic — and largely on American MBA students. The Want Self/Should Self framework assumes a particular cultural psychology of individual deliberation. Honor cultures, shame cultures, collectivist cultures, and high-corruption normalized environments may show different ethical decision architectures. Generalizations like “ethical fading is universal” overstate the evidence base, which is wide in the US and Western Europe but thin in the rest of the world. International deployments of behavioral-ethics interventions should treat the local cultural context as a moderator, not a wrapper.

What’s Really Happening Inside the Brain

Behavioral ethics is descriptive psychology by training, but the underlying neuroscience has caught up enough that the major mechanisms have plausible biological correlates. The picture is partial but converging.

The Should Self versus Want Self distinction maps loosely onto the well-known dual-process distinction between reflective and reactive systems. Joshua Greene’s fMRI work on moral dilemmas shows that deliberative ethical judgments recruit the dorsolateral prefrontal cortex (dlPFC), the brain region associated with effortful executive control, while reactive judgments under time pressure recruit the ventromedial prefrontal cortex (vmPFC) and amygdala, regions tuned to affect and immediate evaluation. The action-phase Want Self looks neurologically like the reactive system; the prediction-phase and recollection-phase Should Self looks like the reflective system. Under cognitive load or time pressure, the reflective system loses bandwidth and the reactive system takes over — which is exactly the operational stress condition under which ethical fading and motivated blindness most reliably fire.

Motivated blindness has a candidate neural mechanism in the way the brain handles incentivized perception. Studies by Florian Engelmann and others have shown that financial conflict of interest shifts attention away from disconfirming evidence at the level of early perceptual processing, not just at the level of conscious deliberation. We do not see the inconvenient data and then choose to ignore it; the data does not reach conscious attention in the first place. This is consistent with broader findings on selective attention from the Chabris and Simons invisible-gorilla experiments — under focused attention, large amounts of perfectly visible information do not register if they are not relevant to the active task.

The slippery slope mechanism has support in fMRI work by Tali Sharot and others showing that the amygdala response to dishonesty habituates with repeated exposure, much like the response to disgust habituates with repeated exposure. The first lie produces a measurable amygdala spike. The second lie produces less. By the tenth lie, the response is statistically indistinguishable from honest behavior. The biology is what makes the path-dependence of the slippery slope real rather than metaphorical.

Outcome bias has a candidate neural mechanism in the dopaminergic prediction-error signal. Rewards that follow our decisions are encoded by the dopamine system in a way that reinforces the decision regardless of its underlying logic. A bad decision followed by a good outcome registers, biologically, as a good decision. The reinforcement is what makes outcome bias so resistant to deliberate correction: by the time the retrospective review fires, the dopamine signal has already done its work, and the reviewer is not so much evaluating the past decision as confirming a stored conclusion about it.

The neuroscience does not change the design implications. It strengthens them. If the mechanisms are biological and the biology is convergent across the population, then design that ignores them is design that fights the wiring. Design that respects them — that reduces operational stress at the moment of ethical decision, that builds in pre-commitment, that surfaces disconfirming evidence by structure rather than by request — works with the wiring instead of against it.

Behavioral Ethics vs Other Theories

Behavioral ethics is not the only descriptive theory of ethical decision-making. It earns its place by what it does that the alternatives don’t.

vs Kohlberg’s Stages of Moral Development

Lawrence Kohlberg’s six-stage developmental theory has been the dominant educational paradigm for ethics for half a century. Its central claim is that moral reasoning matures through identifiable stages, from punishment-avoidance through social-contract reasoning up to universal ethical principles. Behavioral ethics does not deny the developmental story; it argues that the developmental level a person has reached does not predict the behavior they will produce in real ethical situations. People at Kohlberg stage 5 can show ethical fading just as reliably as people at stage 3. The variable that matters in the moment is the situational architecture, not the developmental height of the reasoner.

vs Rest’s Four-Component Model

James Rest’s four-component model (moral sensitivity, moral judgment, moral motivation, moral character) is the closest mainstream competitor and the most useful precursor by a wide margin. Behavioral ethics adopts the architecture — multi-component, descriptive, attention-aware — but goes further in two ways. First, it produces specific mechanisms within each component, not just labels for the components. Second, it foregrounds the failure modes of moral sensitivity (the first component) in a way Rest underweighted; most of the action in behavioral ethics is at the sensitivity stage, where the ethical content of the situation either registers or does not.

vs Gentile’s Giving Voice to Values

Mary Gentile’s Giving Voice to Values framework is the most practical applied alternative. Gentile takes the position that the hard problem is not knowing the right thing but speaking up about it in organizational settings where doing so is costly. Her training program is built around scripts and rehearsal — preparing people to act on values they already hold. Behavioral ethics and Giving Voice to Values are complements rather than rivals. Behavioral ethics describes why the Want Self loses; Giving Voice to Values trains the Should Self to be more capable of speaking up at the moments behavioral ethics warns are dangerous. The fully equipped designer reads both.

vs Cialdini’s Influence and Persuasion Stack

Robert Cialdini’s six principles of influence operate at a different layer. Cialdini’s principles describe the levers persuasion uses to move people; behavioral ethics describes the levers misconduct uses to recruit otherwise ethical people. The two literatures share a vocabulary (reciprocity, social proof, authority, scarcity) but ask different questions. Read together, they expose a useful identity: most of the mechanisms Cialdini documents as engines of compliance are mechanisms behavioral ethics documents as engines of unethical compliance under bad design.

Behavioral Ethics in the Real World

The five mechanisms are easiest to recognize in the corporate failures that have stress-tested the field’s predictions since 2011.

Wells Fargo (2016 and Continuing): The Aggressive-Goal Failure

The Wells Fargo cross-sell scandal — in which roughly 3.5 million unauthorized accounts were opened to meet aggressive cross-sell quotas — is the textbook contemporary case. The diagnosis is straightforward through the behavioral-ethics lens. The quota system created a Want Self for every branch employee that could only succeed by opening accounts customers had not requested. The quota was framed as a business target, not an ethical risk, which is ethical fading at scale. Each individual unauthorized account was small enough to feel like a survival workaround, which is the slippery slope. Branch managers were paid on the quota and were, by structure, motivatedly blind to evidence of the harm. The CFPB enforcement action ultimately spanned multiple years of settlements continuing into 2022 and beyond, and the leadership defense in every depositions phase took the form of “we did not know.” Behavioral ethics says they probably didn’t — not because the evidence wasn’t there, but because the system was designed in a way that made the evidence invisible to the people who needed to see it.

Volkswagen Dieselgate (2015 Onward): Indirect Blindness at Scale

The Volkswagen emissions scandal involved roughly eleven million vehicles fitted with software that detected emissions testing conditions and produced false readings. The engineering work was distributed across multiple teams over several years; no single engineer was responsible for the full system. The indirect-blindness mechanism is visible at every layer: each engineer worked on a defensible technical sub-problem, the management chain saw aggregate emissions outcomes rather than the implementation details, and the corporate leadership saw market share. The aggregate misconduct exceeded what any individual decision-maker would have authorized if presented with the full picture. The Department of Justice settlements eventually totaled over twenty billion dollars in the United States alone.

Boeing 737 MAX (2018-2024): Time Pressure and the Scarcity Trap

The two crashes of Boeing 737 MAX aircraft in 2018 and 2019, which killed 346 people, and the subsequent 2024 door-plug incident have a different shape but the same architecture. Internal Boeing communications released during congressional investigations show engineers expressing concerns about the MCAS flight-control system years before the crashes. The concerns were not suppressed in any obvious way; they simply did not survive the time-pressure environment of a program racing to compete with the Airbus A320neo. Each individual decision — to scope MCAS narrowly, to omit the system from pilot training documentation, to certify under a streamlined process — was defensible on its own terms. The aggregate produced two crashes. Behavioral ethics names this pattern with the clearest scarcity-and-loss vocabulary in the field: when Core Drive 6 (CD6): Scarcity & Impatience (the deadline lever) dominates the operational frame, ethical attention collapses systematically.

FTX and Theranos: Mission as Ethical Camouflage

The FTX collapse in November 2022 and the Theranos verdicts in January 2022 share a behavioral-ethics fingerprint. In both cases, founders with intensely articulated missions — effective altruism in Sam Bankman-Fried’s case, healthcare democratization in Elizabeth Holmes’s case — presided over conduct that, viewed in isolation, would have been recognized as fraud by every employee involved. The mission did not cause the fraud, but it provided the ethical-fading scaffolding that made the fraud frame-able as something else. Bazerman has written specifically about the SBF case as a textbook demonstration of how a high-Epic-Meaning operating environment can become the most permissive environment imaginable for bounded ethicality. The lesson is uncomfortable for any designer building a mission-driven product: the stronger the mission, the more careful the ethical scaffolding around it has to be.

The Elephant in the Room

Behavioral ethics has a problem it does not love to discuss. The same field that names how good people slide into unethical decisions is being used, with increasing sophistication, by the people building the systems that produce the unethical decisions. The techniques that make ethical fading visible to an outside auditor are the techniques that an inside designer can use to engineer ethical fading more reliably. The literature on motivated blindness in auditors is, viewed from the other side, a literature on how to construct an incentive structure that produces motivated blindness on demand.

This is not a hypothetical concern. The dark-pattern research community has documented growing use of behavioral-ethics insights to build interfaces specifically designed to bypass the Should Self at the moment of decision. Subscription traps that exploit time pressure, default settings that exploit slippery-slope acceptance, “anchor account” inducements that exploit endowment-by-the-time-you-notice — the entire dark-pattern toolkit can be read as bounded-ethicality research weaponized into the design of choice architecture.

The honest position for any designer working with these tools is to acknowledge the dual-use nature of the field. Naming this directly is the first defense against being a participant in it. Bazerman and Tenbrunsel themselves have noted, in later writing, that the behavioral-ethics agenda must include defenses against the same mechanisms it documents — not just diagnostics of those mechanisms. The publicity test from the Libertarian Paternalism pillar is the canonical move here: would you be willing to publish, on the front page of your own newsletter, the specific design decision you are about to ship? If the answer is no, the design is using behavioral ethics offensively, not defensively, and the ethics fingerprint matches the cases above more than it matches the field’s stated purpose.

The deeper elephant is that the entire behavioral-ethics field is, itself, an applied program in a discipline (academic ethics) that historically resisted application. Some normative ethicists view the field with suspicion, arguing that descriptive findings about how humans fail at ethics are being used to lower the bar for ethical conduct rather than to raise it. The position is not without merit. If “the Want Self always loses” becomes the operating assumption, the resulting design tolerance for Want Self conduct ratchets upward. Behavioral ethics, used carelessly, can become a sophisticated way of giving up. The discipline that gets the most out of the field is the one that uses descriptive findings to design tighter, not looser, environments.

How to Apply Behavioral Ethics with the Octalysis Framework

Below is the Octalysis Framework with Game Techniques diagram — the eight Core Drives arranged around the octagon, each Drive paired with the techniques that activate it. This is the master reference for everything that follows.

The unique contribution this pillar makes is what I call the Octalysis-Grey-Hat audit — a per-Core-Drive map of where each motivational lever in a system also functions as a bounded-ethicality vector. Every Drive that motivates engagement also recruits behavior the Want Self would not have chosen if the Should Self had been in the room. Naming the vector for each Drive is what lets a designer audit a live system for ethical risk before the risk hits the front page.

Core Drive 1: Epic Meaning & Calling — The Cause-Hijack Vector

Core Drive 1 (CD1): Epic Meaning & Calling is the Drive that lets people sustain effort for outcomes that benefit something larger than themselves. It is the Drive behind Wikipedia volunteers, MrBeast’s PhilanthropyU production teams, and every faith-based social network. The bounded-ethicality vector is what I call cause-hijack: when the mission is loud enough, it begins to function as a moral exemption for conduct that, on its own merits, would have been recognized as misconduct. FTX is the textbook case. Theranos is another. The internal narrative is “we are too important to fail under conventional ethical constraints,” which is exactly the framing ethical fading produces at the organizational scale. Defensive design: install pre-commitment ethical guardrails that fire before the mission narrative is loud enough to override them. Once the cause-hijack frame is active, in-the-moment defenses do not work.

Core Drive 2: Development & Accomplishment — The Goal-Setting Trap

Core Drive 2 (CD2): Development & Accomplishment runs on visible progress, milestones, and aggressive targets. It is the most-used Drive in commercial gamification by a wide margin. The bounded-ethicality vector is the goal-setting trap documented by Ordóñez, Schweitzer, Galinsky, and Bazerman in their 2009 paper Goals Gone Wild. Aggressive numerical goals systematically generate unethical behavior at the margin — not because the goal-setters intend to corrupt anyone, but because the marginal cost of cheating drops sharply when the target is just out of reach. The Sears auto-repair quota in the early 1990s produced fabricated repair recommendations. The Wells Fargo cross-sell quota produced unauthorized accounts. Every aggressive CD2 system carries this risk, and the higher the quota, the more reliably the risk materializes. Defensive design: pair every aggressive numerical target with an ethical-quality gate that has independent budget and is empowered to slow the quota when the cheating tail starts appearing.

Core Drive 3: Empowerment of Creativity & Feedback — The Creative Cheating Engine

Core Drive 3 (CD3): Empowerment of Creativity & Feedback drives the design-and-iterate loop that makes great products possible. The bounded-ethicality vector is the creative-cheating engine documented by Francesca Gino and Dan Ariely in their 2012 paper. Creative people, on average, cheat more in laboratory tasks than less creative people, and the mechanism is exactly what makes them creative: they generate justifications faster. The same fluency that produces novel solutions produces novel rationalizations. The implication for design is that CD3-heavy environments — agencies, R&D groups, startups in stealth — carry a disproportionate ethical-fading risk per unit of activity, and the standard intervention (rules and policies) is the wrong intervention because the creative environment is selected for routing around rules and policies.

Core Drive 4: Ownership & Possession — The Endowment-Conflict Vector

Core Drive 4 (CD4): Ownership & Possession is the Drive behind virtual goods, customization, equity grants, and any system that lets a user build something they consider theirs. The bounded-ethicality vector is the endowment-conflict vector documented across the motivated-blindness literature. Once a person owns a position — financial, intellectual, emotional — their attention bends in the direction of preserving that position. Auditors paid by audited firms cannot see the inconvenient evidence. Analysts whose investment thesis depends on a particular company’s success cannot see the disconfirming data. Founders who own a substantial equity stake in their startup cannot see the early signs that the business is not going to work. Defensive design: structurally separate the person who benefits from a decision from the person who makes the decision, especially at the most consequential moments. The auditor-independence example is the canonical instance and most other CD4-driven motivated-blindness cases have the same structural fix.

Core Drive 5: Social Influence & Relatedness — The Conformity-to-Wrong Vector

Core Drive 5 (CD5): Social Influence & Relatedness is the Drive behind peer comparison, group identity, and the entire engagement-loop of social products. The bounded-ethicality vector is conformity-to-wrong — a generalization of the classic Asch line-judgment finding into ethical territory. When peers normalize an unethical practice, individual judgment slides toward the peer norm, even for individuals who, asked privately, would describe the practice as unethical. The Volkswagen engineering teams are the cleanest example: each engineer saw colleagues working on the cheat device and each engineer concluded, reasonably, that the cheat device must be permissible if every other engineer was working on it. Defensive design: install ethical-reporting channels that operate independently of the peer network. Solomon Asch’s own follow-up work showed that the presence of even one dissenter collapses the conformity effect, which means the design move is to guarantee that a dissenter can exist without social cost.

Core Drive 6: Scarcity & Impatience — The Time-Pressure Engine

Core Drive 6 (CD6): Scarcity & Impatience drives every deadline-based mechanic, every limited-time offer, every quarter-end push. The bounded-ethicality vector is the time-pressure engine documented in Darley and Batson’s 1973 Good Samaritan study and replicated across the field. Moral attention collapses systematically under time pressure. Seminary students late to a lecture step over a victim that students with time to spare stop to help. The mechanism scales: Boeing engineers under program-deadline pressure approved certifications they would have flagged with more time. Wells Fargo branch employees under daily-quota pressure opened accounts they would not have opened at month-start. Defensive design: build deliberate ethical-decision slack into the highest-stakes operational moments. Counter-intuitively, the most ethically resilient systems are the ones that explicitly slow down at the moments the operational economics most demand speed.

Core Drive 7: Unpredictability & Curiosity — The Outcome-Bias Engine

Core Drive 7 (CD7): Unpredictability & Curiosity is the Drive behind every gambling mechanic, every variable-reward schedule, and every retrospective-narrative dynamic. The bounded-ethicality vector is the outcome-bias engine. Decisions made under uncertainty are judged retrospectively by their outcomes rather than by the quality of the reasoning behind them. The CD7 environment makes this worse, because the variance is high enough that bad decisions reliably produce good outcomes some of the time, and good decisions reliably produce bad outcomes some of the time. The most dangerous executives are the ones with prior high-risk gambles that paid off, because their internal narrative tells them their judgment is sound. Defensive design: judge decisions by their process at the moment they were made, not by their outcomes. Independent process-quality reviews, run before the outcome is known, are the only intervention that systematically corrects outcome bias.

Core Drive 8: Loss & Avoidance — The Cover-Up Engine

Core Drive 8 (CD8): Loss & Avoidance is the Drive that activates when something we have is threatened. It is the most powerful single Drive in the Octalysis Framework by short-term motivational intensity. The bounded-ethicality vector is the cover-up engine, sometimes called the Watergate principle: the largest unethical decisions in most scandals are not the original misconduct but the subsequent decisions to conceal it. The CD8 activation behind the cover-up is more intense than the CD8 activation behind the original act, because by the time the cover-up is at stake, the actor has more to lose. This is why so many scandals follow the same arc — a manageable initial violation, an unmanageable concealment, and a catastrophic exposure. Defensive design: structure the post-mistake recovery path so that the cost of admitting the original violation is dramatically lower than the cost of concealing it. Most organizations do this exactly backwards; their concealment infrastructure (legal, PR, internal politics) is mature while their admission infrastructure is improvised, which guarantees the cover-up wins.

The 6-Step Behavioral Ethics × Octalysis Audit

The per-Drive map is diagnostic. The following six-step audit turns it into a defensive design practice that any commercial system can run, including consumer apps, internal organizational designs, and policy interventions.

Step 1: Map every active Core Drive in the system to its bounded-ethicality vector. Most systems are running three or four Drives simultaneously. Each one carries its own vector. Naming them is the first move, because what is unnamed cannot be audited.

Step 2: Pre-mortem the Want Self under operational stress. For each named vector, ask: under what realistic operational pressure does the Want Self take over? When does the deadline get tight enough, the quota aggressive enough, or the peer norm normalized enough that the Should Self stops being in the room? The answer should be specific. Specific is auditable.

Step 3: Run a sludge-versus-grease audit on ethical attention paths. Friction that protects ethical decision-making (deliberate slowdowns, mandatory review, structural separation) is grease, not sludge. Friction that protects unethical decision-making (delayed admission paths, confusing reporting channels, ambiguous escalation rules) is sludge that must be removed. Most organizations are calibrated exactly backwards.

Step 4: Run a goal-setting toxicity check on every CD2 mechanic. Aggressive numerical targets are the single most reliable bounded-ethicality generator in commercial systems. Any quota that exceeds the median performer’s prior-period production by more than a documented amount needs a paired ethical-quality gate. If the gate cannot be funded, the quota should not be set.

Step 5: Install an outcome-bias firewall on retrospective reviews. Every consequential decision should be reviewed for process quality before its outcome is known and that review should be sealed and revealed at the retrospective alongside the outcome data. Without this firewall, the dopaminergic prediction-error signal will rewrite the organizational memory of the decision, and the wrong lessons will compound.

Step 6: Apply the publicity test. Would you publish the specific design decision on the front page of your own newsletter? If not, the design is using behavioral ethics offensively. The publicity test is the cheapest, fastest defense against accumulated bounded-ethicality drift in any commercial system.

Practical Steps for Designers and Operators

1. Run the per-Drive vector map on every system you ship. The diagnostic takes thirty minutes per product. The audit becomes part of the design ritual, not a once-a-year ethics review by people who do not touch the product.
2. Pair every aggressive CD2 quota with an ethical-quality gate that has independent budget and is empowered to slow the quota. The pairing is non-optional. Quotas without paired gates produce Wells Fargo by structural law.
3. Structurally separate decision authority from financial benefit at every consequential moment. The auditor-independence pattern generalizes. If the person who benefits from a decision is also making it, motivated blindness is mathematically guaranteed.
4. Install a sealed-process-review firewall on every consequential decision. Review the decision quality before the outcome is known and reveal both at retrospective. The firewall stops the dopaminergic rewrite.
5. Make admission cheaper than concealment. Build admission infrastructure that is at least as mature as your concealment infrastructure. The CD8 cover-up vector is decisive when the admission path is improvised and the concealment path is institutional.
6. Run the publicity test before every consequential design change. If the change cannot survive front-page exposure of the specific decision, do not ship it. The test takes one minute and catches more bounded-ethicality drift than any ethics committee.
7. Audit your own use of these tools every six months. The dual-use nature of behavioral ethics means that the people who know it best are also the people most equipped to abuse it. Self-application of the publicity test, on your own behavioral-ethics work, is the discipline that distinguishes the field from its dark-pattern shadow.

Behavioral Ethics Was the Beginning, Not the End

Bazerman and Tenbrunsel framed the field as descriptive psychology in the service of better organizational design. The fifteen years since Blind Spots have shown both the strength and the limit of that framing. The strength is that the diagnostic vocabulary works: every major corporate ethical failure since 2011 reads cleanly through the bounded-ethicality lens, and the lens has predictive value for new failures as they emerge. The limit is that descriptive psychology, on its own, does not produce reliable interventions. The field is still maturing toward a science of defensive design, and most of the practical work is being done outside the academy — in regulatory rulemaking, in technical standards, in the dark-pattern research community, and in the small number of in-house ethics teams that have been resourced to do the work.

For a behavioral designer, the field is best understood as a Grey Hat companion to Octalysis. Every Core Drive that motivates engagement carries a bounded-ethicality vector. The job is not to choose between motivation and ethics — that’s a false choice that produces the under-designed ethics products that never get used. The job is to build the motivation and the ethical defense in the same architecture, knowing that the audience will be the Want Self at the moment of decision and that the Should Self will not be in the room. The systems that survive front-page exposure are the ones whose designers wrote for the Want Self while protecting the Should Self by structure.

If you have read this far, the next move is to run the six-step audit on whatever system you are responsible for. Pick the highest-leverage CD2 quota in your current operation, find the paired ethical-quality gate, and verify that the gate has actual budget. If it does not, schedule the conversation that funds it. That single move — one quota, one paired gate — is more behavioral-ethics work than most organizations do in a year, and it is exactly the work that catches the next Wells Fargo before it ships.

Frequently Asked Questions

What is behavioral ethics in one sentence?

Behavioral ethics is the descriptive study of how good people end up making unethical decisions they would have rejected in advance, organized around the psychological mechanisms that operate below conscious awareness.

Who founded the field?

Max Bazerman at Harvard Business School and Ann Tenbrunsel at Notre Dame are the most-cited founders, with their 2011 book Blind Spots: Why We Fail to Do What’s Right and What to Do about It serving as the field’s definitive synthesis. The broader research program draws on earlier work by Herbert Simon (bounded rationality), David Messick, James Rest, and others.

What is bounded ethicality?

Bounded ethicality is the umbrella term for the systematic psychological constraints on ethical reasoning that produce unethical behavior even from people who would describe themselves, accurately, as ethical. It is the deliberate analogue to Simon’s bounded rationality and operates through five distinct mechanisms.

What are the five mechanisms of bounded ethicality?

The five mechanisms are ethical fading (the ethical content of a decision recedes from awareness), motivated blindness (failure to see information that, if seen, would be inconvenient), indirect blindness (less harsh judgments of misconduct done through intermediaries), the slippery slope (small violations compound invisibly into large ones), and outcome bias (judgments of past ethics depend on outcomes, not process).

How does the Want Self differ from the Should Self?

The Should Self is the deliberative, principled actor we are in the prediction and recollection phases of ethical decisions. The Want Self is the reactive, situationally-driven actor we are in the action phase. The two are inhabited by different cognitive systems — one reflective, one reactive — and the action-phase Want Self is the one that actually controls behavior under operational pressure.

Why don’t ethics training programs work?

Most ethics training targets the Should Self in a quiet conference room. The Want Self, who actually makes the decisions under operational pressure, is not in the room and is not addressable by the training format. Effective interventions change the moment of decision — the choice architecture, the time pressure, the peer norm — rather than the principles of the decider.

How does behavioral ethics relate to Cialdini’s principles of influence?

Most of the mechanisms Cialdini documents as engines of compliance are mechanisms behavioral ethics documents as engines of unethical compliance under bad design. The two literatures share a vocabulary and ask different questions; read together, they expose the dual-use nature of every persuasion principle.

What is the publicity test?

The publicity test asks whether you would be willing to publish, on the front page of your own newsletter, the specific design decision you are about to ship. It is the cheapest, fastest defense against accumulated bounded-ethicality drift in any commercial system. Most accumulated misconduct fails the test even when its individual steps do not.

How does the Octalysis Framework integrate with behavioral ethics?

Every Octalysis Core Drive that motivates engagement is also a bounded-ethicality vector. The Octalysis-Grey-Hat audit maps each Drive to its specific vector (CD2 to goal-setting traps, CD6 to time pressure, CD8 to the cover-up engine, and so on) and then runs a six-step defensive design pass on each named vector. The integration is what lets a designer audit a live system for ethical risk before the risk hits the front page.

Is behavioral ethics being used for dark patterns?

Yes, and acknowledging this directly is part of working with the field responsibly. The same techniques that make ethical fading visible to an outside auditor are the techniques an inside designer can use to engineer ethical fading more reliably. The publicity test is the canonical defense, and self-application of the test on your own behavioral-ethics work is the discipline that separates the field from its dark-pattern shadow.

References

1. Bazerman, M. H., & Tenbrunsel, A. E. (2011). Blind Spots: Why We Fail to Do What’s Right and What to Do about It. Princeton University Press. (The field’s definitive book; mandatory reading for any commercial application.)
2. Tenbrunsel, A. E., & Messick, D. M. (2004). Ethical fading: The role of self-deception in unethical behavior. Social Justice Research, 17(2), 223-236. (Canonical ethical-fading paper.)
3. Moore, D. A., Tetlock, P. E., Tanlu, L., & Bazerman, M. H. (2006). Conflicts of interest and the case of auditor independence: Moral seduction and strategic issue cycling. Academy of Management Review, 31(1), 10-29.
4. Shu, L. L., Gino, F., & Bazerman, M. H. (2011). Dishonest deed, clear conscience: When cheating leads to moral disengagement and motivated forgetting. Personality and Social Psychology Bulletin, 37(3), 330-349.
5. Gino, F., & Bazerman, M. H. (2009). When misconduct goes unnoticed: The acceptability of gradual erosion in others’ unethical behavior. Journal of Experimental Social Psychology, 45(4), 708-719. (Canonical slippery-slope demonstration.)
6. Baron, J., & Hershey, J. C. (1988). Outcome bias in decision evaluation. Journal of Personality and Social Psychology, 54(4), 569-579.
7. Ordóñez, L. D., Schweitzer, M. E., Galinsky, A. D., & Bazerman, M. H. (2009). Goals gone wild: The systematic side effects of overprescribing goal setting. Academy of Management Perspectives, 23(1), 6-16. (The CD2 goal-setting trap canonical.)
8. Gino, F., & Ariely, D. (2012). The dark side of creativity: Original thinkers can be more dishonest. Journal of Personality and Social Psychology, 102(3), 445-459.
9. Darley, J. M., & Batson, C. D. (1973). From Jerusalem to Jericho: A study of situational and dispositional variables in helping behavior. Journal of Personality and Social Psychology, 27(1), 100-108. (The Good Samaritan time-pressure canonical.)
10. Asch, S. E. (1956). Studies of independence and conformity: A minority of one against a unanimous majority. Psychological Monographs, 70(9), 1-70.
11. Greene, J. D., Sommerville, R. B., Nystrom, L. E., Darley, J. M., & Cohen, J. D. (2001). An fMRI investigation of emotional engagement in moral judgment. Science, 293(5537), 2105-2108. (Neural correlates of deliberative vs. reactive moral judgment.)
12. Garrett, N., Lazzaro, S. C., Ariely, D., & Sharot, T. (2016). The brain adapts to dishonesty. Nature Neuroscience, 19(12), 1727-1732. (Amygdala habituation in the slippery slope.)
13. Rest, J. R. (1986). Moral Development: Advances in Research and Theory. Praeger. (Four-component model precursor to behavioral ethics.)
14. Kohlberg, L. (1981). The Philosophy of Moral Development: Moral Stages and the Idea of Justice. Harper & Row.
15. Gentile, M. C. (2010). Giving Voice to Values: How to Speak Your Mind When You Know What’s Right. Yale University Press.
16. Simon, H. A. (1957). Models of Man. Wiley. (The bounded-rationality canonical that bounded ethicality deliberately mirrors.)
17. Sunstein, C. R. (2022). Sludge: What Stops Us from Getting Things Done and What to Do about It. MIT Press. (Companion architectural lens to behavioral ethics.)
18. Tenbrunsel, A. E., & Smith-Crowe, K. (2008). Ethical decision making: Where we’ve been and where we’re going. Academy of Management Annals, 2(1), 545-607.
19. Chugh, D., Bazerman, M. H., & Banaji, M. R. (2005). Bounded ethicality as a psychological barrier to recognizing conflicts of interest. In Conflicts of Interest: Challenges and Solutions in Business, Law, Medicine, and Public Policy, edited by D. A. Moore et al. Cambridge University Press.
20. Chou, Y.-k. (2015). Actionable Gamification: Beyond Points, Badges, and Leaderboards. Octalysis Media. (The Octalysis Framework canonical, including the Grey Hat / White Hat axis.)
21. Bazerman, M. H. (2020). Better, Not Perfect: A Realist’s Guide to Maximum Sustainable Goodness. HarperBusiness. (Bazerman’s update incorporating subsequent field evidence.)

Related Reading

• Libertarian Paternalism: The Ethics of Choice Architecture — The companion ethics frame for any behavioral-design intervention.
• Moral Disengagement (Bandura): The Eight Mechanisms — Bandura’s parallel taxonomy of self-exoneration; the closest sibling literature to bounded ethicality.
• Sludge: The Frictions That Stop Beneficial Action — Sunstein’s frame for the friction-side of the architecture.
• Precaution Adoption Process Model — The stage-matched intervention map that pairs naturally with behavioral-ethics audits.
• The Complete Octalysis Framework — The motivational core that the Grey Hat audit is run against.

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?