Why useful systems begin with responsibility

“Some things are in our control and others not.”
— Epictetus, Enchiridion, §1 [1]

A useful system does not merely assign responsibility. It gives a person a fair chance to carry it.

Consider a familiar meeting.

A deadline has been missed. The work moved from strategy to design, from design to operations, and from operations to a partner outside the organization. Every team completed something. Nobody protected the whole outcome. By the time the failure became visible, each person could explain—with some accuracy—why the decisive moment belonged to someone else.

The meeting ends with familiar remedies: communicate earlier, collaborate better, take more ownership.

These are morally satisfying instructions. They are also often poor system design.

“Take ownership” is easy to say when the outcome is visible, authority is matched to the task, information arrives in time, and escalation is safe. It becomes theatre when a person is held answerable for a result they could not see, alter, or challenge.

That distinction matters in a classroom, a hospital, a start-up, a university, and a multinational firm. It matters even more as intelligent tools become participants in workflows without becoming moral participants in their consequences.

This first Iesue therefore begins with a deceptively simple claim:

A useful system begins by asking what responsibility a person is carrying—and then designing the conditions that make responsible action possible.

Responsibility is not blame

Organizations frequently confuse responsibility with the location at which blame should stop.

Blame looks backward. It asks who can be attached to the failure. Responsibility is wider. It asks who must notice, decide, act, communicate, and repair—before the failure becomes inevitable.

The distinction is not an argument against individual accountability. People can act carelessly, conceal information, avoid difficult conversations, or neglect standards they understood perfectly well. Character matters. The Mirror comes before public trust.

But character never works on empty ground.

Epictetus begins the Enchiridion by separating what belongs to our agency from what does not. The insight is usually presented as private philosophy: govern your judgment and actions rather than demanding control over reputation, property, or other people. It is also a powerful question for organizational design.

What, exactly, have we placed within this person’s control?

If someone owns an outcome but cannot obtain the necessary information, change the process, question the premise, refuse unsafe work, or reach the person who can intervene, the organization has not created responsibility. It has created exposure.

Accountability without corresponding agency is not rigor. It is a sophisticated form of blame.

The system is already teaching

Every system teaches people what responsibility means inside it.

A dashboard teaches which outcomes count. A meeting teaches whose uncertainty is welcome. A deadline teaches what may be sacrificed. A promotion teaches which behaviour the institution truly values. A course rubric teaches what kind of thinking deserves attention. An unanswered warning teaches whether speaking up was worth the cost.

This is why the National Academies’ landmark patient-safety report, To Err Is Human, moved the discussion beyond exhorting professionals to “be more careful.” Its design principle was plain:

“Errors can be prevented by designing systems that make it hard for people to do the wrong thing and easy for people to do the right thing.” [2]

Health care is a high-stakes domain, and lessons from it should not be transplanted carelessly into every office or classroom. But the underlying logic travels well: when a pattern of failure recurs across otherwise capable people, another motivational speech is weak medicine. The pattern is information about the ground.

Research on work design makes a related point. Hackman and Oldham’s job-characteristics model links autonomy with an experienced sense of responsibility for outcomes, while feedback helps people know the results of their work. [3] Responsibility is therefore not produced by a job title alone. It is experienced when a person can exercise meaningful discretion and encounter intelligible consequences.

Research on coordination pushes the argument further. Okhuysen and Bechky propose that mechanisms such as plans, routines, schedules, meetings, and roles matter because they help create three conditions: accountability, predictability, and common understanding. [4] A process is useful not because it exists, but because it helps interdependent people understand who carries what, what others are likely to do, and what situation they are jointly facing.

That is a higher standard than administrative neatness.

The six conditions of responsible work

Before asking a person to “own” an outcome, inspect six conditions.

1. A visible outcome

Can the person describe what must be protected—not merely which task must be completed?

Tasks encourage local completion. Responsibility requires sight of the consequence. “Send the report” is a task. “Give the decision-maker an accurate basis for acting by Thursday” is an outcome.

2. Matched authority

What can the person decide without seeking permission? What may they stop? Which trade-offs remain theirs?

Autonomy does not mean isolation or unlimited discretion. It means that the authority granted is proportionate to the responsibility assigned.

3. Timely information

What must the person know, and when must they know it?

Information delivered after the decision is documentation, not support. Useful systems place evidence near the moment of judgment.

4. Explicit handoffs

Where does responsibility move—and how do both sides know that it has moved?

“I sent it” is not the same as “the next owner received, understood, and accepted it.” Handoffs require a visible change in custody.

5. Permission to surface trouble

Can someone say “I do not know,” “this assumption is wrong,” or “we should stop” without first calculating the interpersonal cost?

Amy Edmondson defined team psychological safety as “a shared belief held by members of a team that the team is safe for interpersonal risk taking.” [5] Her study connected that climate with team learning behaviour. Psychological safety is not freedom from standards. It is the condition under which people can reveal the information that responsible standards require.

6. Feedback and repair

Can the person see what happened, learn from it, and participate in repair?

A system that records mistakes but does not return usable information to the people doing the work creates surveillance, not learning.

The responsibility test
Do people have sight of the outcome, authority to act, information in time, explicit handoffs, permission to surface trouble, and a route to repair?

The classroom is an organization too

The classroom offers a particularly honest test of this argument.

Professors often say they want students to take responsibility for their learning. Then we design a course in which the important decisions remain invisible, the standards appear only at the moment of judgment, practice is scarce, and feedback arrives after it can be used.

In that environment, “take responsibility” may mean little more than “absorb the consequences of a design you did not understand.”

A demanding education should ask more of students. It should also make the route into demanding work visible. Students need opportunities to interpret criteria, practise difficult thinking, receive information, revise, and gradually carry more consequential choices. The instructor remains responsible for the quality of the ground while refusing to walk the student’s path on their behalf.

This is supported rigor: meaningful standards, genuine agency, and conditions worthy of the effort being requested.

The same principle changes group work. A group grade does not create shared responsibility by itself. Students need an intelligible common outcome, differentiated roles, ways to make progress visible, explicit handoffs, and a process for raising a problem before resentment becomes the only available form of feedback.

When those conditions are absent, the most conscientious student often carries the invisible weight. The least informed student discovers the failure at the grade. Neither learns much about responsible collaboration.

A classroom is not rigorous because consequences are severe. It is rigorous when students can see, practise, and increasingly carry the responsibilities that capable work requires.

Where the argument meets resistance

There is a danger in every systems argument: the system can become an all-purpose alibi.

If every failure is attributed to culture, incentives, workload, technology, or leadership, individual agency disappears. People become plants waiting for better soil. That is not the Yes-Way.

The Mirror still asks what was yours to govern. The Weight still asks whom your action affected. The Path still requires a decision that no institution can make entirely for you.

The better sequence is diagnostic rather than ideological:

Was the outcome clear?
Did the person possess proportionate authority and timely information?
Were the handoffs and escalation route workable?
Could concerns be raised without unreasonable penalty?
If those conditions were present, did the person exercise responsible judgment?

Sometimes the answer to the fifth question is no. Accountability is then necessary. But asking it first allows a badly designed system to keep consuming good people while calling the damage a character problem.

Meta-analytic research has found a negative relationship between role ambiguity and job performance. [6] That does not mean every uncertainty can or should be eliminated. Leadership and entrepreneurship often require action under ambiguity. The relevant distinction is between productive uncertainty in the problem and preventable ambiguity about responsibility.

We should not promise people certainty. We should tell them what judgment remains theirs.

The next ground: responsibility with AI

This distinction is becoming more urgent.

Artificial intelligence can now draft, rank, recommend, summarize, predict, and sometimes act. It can participate in a workflow without experiencing duty, reputation, remorse, or the obligation to repair. The danger is not only that a system may produce a poor answer. It is that organizations may allow responsibility to dissolve across the people, vendors, models, interfaces, and approval steps surrounding that answer.

“The AI suggested it” is the new version of “I thought someone else owned it.”

The United States National Institute of Standards and Technology’s AI Risk Management Framework therefore treats governance as a continuing organizational function. It calls for documented roles, clear lines of communication, differentiated responsibilities for human–AI configurations, and accountability structures in which appropriate people are empowered and trained to manage risk. [7]

The enduring question is not whether a human clicked the final button. It is whether a responsible person or group could understand the decision, challenge it, intervene in time, and repair its consequences.

For any AI-assisted workflow, five responsibilities should be named before deployment:

Who verifies the relevant inputs?
Who may override or stop the system?
Who monitors consequences after use?
Who receives a concern or appeal?
Who repairs harm when the workflow fails?

Automation can redistribute work. It cannot absolve an institution of answerability.

A tool may carry out an action. Only a person or institution can carry the responsibility for what that action does.

At four levels

For the person

Do not accept “ownership” as praise until you understand the authority, information, and consequence attached to it. Ask what remains within your control, what must be escalated, and what evidence will tell you whether the work helped.

For the classroom

Before assessing a complex performance, make the responsibility learnable. Show the standard, model the thinking, provide a lower-risk attempt, return feedback while it can still change the work, and give students progressively greater choice.

For the leader

When failure repeats, resist asking who needs another reminder. Map the outcome, authority, information, handoffs, voice, and repair loop. Then decide whether the problem is capability, conduct, coordination, or design.

For the institution

Audit what the system asks people to carry against what it allows them to see and change. Pay particular attention to boundary roles: coordinators, programme staff, executive assistants, adjunct faculty, nurses, customer-support teams, and others who frequently hold consequences without corresponding authority.

The practice: a Ground Audit for responsibility

Choose one recurring responsibility—not an entire department. Use a missed deadline, student project, client handoff, hiring decision, assessment, or AI-assisted workflow.

Write the responsibility in one sentence:

[Person or role] protects [outcome] for [people affected] by [decision or action].

Then ask:

Outcome: Can the person see the consequence, or only the task?
Authority: Which necessary decision can they not make?
Information: What arrives too late, in the wrong form, or not at all?
Handoff: Where could two people reasonably believe the other now owns it?
Voice: What truth is socially or professionally expensive to say?
Repair: How does the system learn after the outcome is known?

Do not redesign everything. Change the smallest condition that would make responsible action more likely and observable. Then watch what happens.

That is the beginning of useful system design: not a new platform, another slogan, or a thicker process manual. A clearer relationship between weight and agency.

The Mirror develops the person. The Weight gives capability a purpose. The Ground makes responsible contribution possible. The Path preserves the judgment no system should take away.

The question worth carrying
Where are you asking someone to carry an outcome while withholding the authority, information, or permission needed to protect it?

Notes and sources

Epictetus. Enchiridion, §1, translated by Elizabeth Carter. Internet Classics Archive, Massachusetts Institute of Technology. Read the text. Translation wording varies by edition.
Institute of Medicine. To Err Is Human: Building a Safer Health System. Washington, DC: National Academies Press, 2000, Preface, p. ix. https://doi.org/10.17226/9728. The quotation is used here as a system-design principle; the report itself concerns patient safety.
Hackman, J. Richard, and Greg R. Oldham. “Motivation through the Design of Work: Test of a Theory.” Organizational Behavior and Human Performance 16, no. 2 (1976): 250–279. https://doi.org/10.1016/0030-5073(76)90016-7.
Okhuysen, Gerardo A., and Beth A. Bechky. “Coordination in Organizations: An Integrative Perspective.” Academy of Management Annals 3, no. 1 (2009): 463–502. https://doi.org/10.5465/19416520903047533.
Edmondson, Amy. “Psychological Safety and Learning Behavior in Work Teams.” Administrative Science Quarterly 44, no. 2 (1999): 350–383. https://doi.org/10.2307/2666999. The original field study involved 51 work teams in one manufacturing company; later research has examined the construct in other settings.
Tubre, Travis C., and Judith M. Collins. “Jackson and Schuler (1985) Revisited: A Meta-Analysis of the Relationships Between Role Ambiguity, Role Conflict, and Job Performance.” Journal of Management 26, no. 1 (2000): 155–169. https://doi.org/10.1177/014920630002600104.
Tabassi, Elham. Artificial Intelligence Risk Management Framework (AI RMF 1.0). NIST AI 100-1. National Institute of Standards and Technology, 2023. https://doi.org/10.6028/NIST.AI.100-1. See especially the Govern function and its provisions on accountability, roles, communication, and human–AI configurations. NIST notes that AI RMF 1.0 is under revision; the cited version remains the published 2023 framework.

A note on the argument

The opening meeting is a composite, not a report of a particular organization. The Yes-Way propositions in this essay—especially the six conditions of responsible work and the four-level application—are my synthesis and interpretation of the cited philosophical, organizational, educational, and governance sources. They should not be read as claims that any single study establishes the complete framework.