Explore the lesson here. Canvas holds your check-ins, submissions, feedback, and progress.

← Learning path

Unit 11 of 12

Make delivery dependable

Who owns quality, exceptions, updates, and continued use?

Learn

By the end of this unit

  • Create operating instructions and a handoff that another person can follow.
  • Respond to changed information or a failure without giving AI unbounded authority.

Planned guided work including the short textbook read-and-apply: 77–127 minutes, plus other project work. This is a design estimate, not measured learner time.

How to read this lesson

The labels describe how to use a passage. Color does not prove a claim is true or that a response is correct.

  • Explain — Understand the idea.
  • Try / predict — Make a choice or prediction before the reveal.
  • Evidence — Inspect the named source or supplied data.
  • Worked reasoning — Compare the steps after your own attempt.
  • Caution — Check a limit, assumption or unsupported conclusion.

Each figure has its own local legend for the data. Blue selection outlines mark the current selection; other outlines are explained by each figure’s local legend. Selection is not a correctness judgment.

Explain

Responsibility continues after the first useful output

Dependable delivery requires more than a successful demonstration. Sources change, users forget instructions, unusual requests arrive, and a busy reviewer can become the bottleneck. A service needs someone responsible for detecting those conditions and deciding what happens next. A personal tool, community project, creative service, or organizational solution still needs a responsible owner.

Imagine assigning AI three bounded roles: organize permitted facts, prepare a draft, and highlight possible unsupported claims. These roles may help one operator. Asking another AI to check the first does not create independent ground truth; both can miss the same source condition. The responsible person checks consequential claims against the approved evidence and controls release.

ReplyKit operating rules for a controlled service
ConditionRequired actionOwner
Ordinary complete inputPrepare, source-check, approve, and record the reply.Coordinator and reviewer
Missing source or policy versionPause drafting and obtain the current approved source.Source owner
Exception or unauthorized promiseHold release; give the owner an evidence-based handoff.Reviewer and operations owner
Reviewer capacity exceededQueue, narrow, or decline work; do not bypass the check.Service owner
Tool unavailableUse the documented manual template and checks.Coordinator
Policy or tool changesVersion the change and retest affected ordinary and exception cases.Maintainer

Define service measures that inform action: acceptable first-pass outputs, material errors, complete effort, time waiting for review, rework, and continued use. A high output count can hide an expanding queue. A support boundary should state what is included, who handles problems, and when the owner reviews continued use. Avoid unlimited promises that the course project cannot maintain.

A handoff is a transfer of the ability to operate, not just the transfer of a file. The recipient needs the current version, purpose, inputs, source, ordinary procedure, exception rule, access location, limitations, and maintenance contact or role. A classroom version can name a role and explicitly state that no ongoing support is provided.

See the decision

A faster draft can build a review queue

Try / predict

Predict what accumulates when 30 requests arrive but only 24 can be reviewed each day.

Supplied case

New fictional five-working-day capacity case, starting with an empty queue. Each reply needs 5 minutes of review. It illustrates operations; it is not observed partner performance.

Thirty arrivals and 24 reviews per working day
DayArrivalsReviewedEnd-of-day queue
130246
2302412
3302418
4302424
5302430
Worked reasoning

Over five days,150 arrive,120 are reviewed, and 30 wait. An additional 30 minutes of review capacity per day handles all 30 arrivals; at $24/hour this adds $60 of assigned staff-time value over the five-day scenario. It is cash expense only if additional time is paid. This is distinct from the lesson’s $60 monthly support assumption. Limiting intake to 24 avoids an internal queue but defers 30 requests over five days; demand has not disappeared.

Use it in your project: Identify the scarce human step, its capacity, what happens when demand exceeds it, and who can change the service promise.

Required textbook reading and application · about 12 minutes

Read, then apply one idea

Read only the passages listed here. Mathematical derivations, algorithms, code, and end-of-chapter exercises are not assigned. The time estimate includes the brief passage and one connection to your existing project update; reading the passage alone usually takes much less time. It is a planning estimate, not a timer.

18.8: Governance and Regulation

David L. Poole and Alan K. Mackworth. Artificial Intelligence: Foundations of Computational Agents. 3rd edition. Cambridge University Press, 2023.

Read: Start at the paragraph beginning "Governance covers government legislation and regulation". Read from there through the discussion of ethics washing and regulatory capture at the end of section 18.8. Stop before navigation. The earlier country-specific and draft-law descriptions are historical and are not assigned as current legal guidance.

Why this unit: Introduces organizational accountability and explains why a human-oversight label alone does not establish safety.

Course connection: Dependable delivery needs a named owner, scheduled checks, a way to report failures, and a working fallback. Human review helps only when the reviewer has the information, skill, time, and authority to act. The original operating checklist translates these general principles into project-specific responsibilities; it is not a legal compliance certification.

Apply one idea: Choose one prompt for the brief connection in your existing project update.

  • Name the owner, a check they will perform, a failure that triggers pausing the solution, and the fallback users can actually follow. State when the next review will happen.
Read this textbook section here

The official page below may include material before or after your assigned selection. The reader does not automatically scroll or crop to that selection.

Assigned selection — 18.8: Governance and Regulation: Start at the paragraph beginning "Governance covers government legislation and regulation". Read from there through the discussion of ethics washing and regulatory capture at the end of section 18.8. Stop before navigation. The earlier country-specific and draft-law descriptions are historical and are not assigned as current legal guidance.

Official textbook section ↗

Add a brief connection to your existing project update: identify the book and section, explain one relevant idea in your own words, and show the decision it informs. This is part of the same update, not another assignment.

If the official reader is unavailable or inaccessible, use the original lesson explanation and worked example, apply the same question, and label that access alternative honestly. Report the problem through the course support route; do not claim to have read an unavailable text.

Optional further reading

The AI Fluency Framework — short guide

Open the short guide here. Identify one decision about delegation, one instruction, one output check, and one responsibility in your task. No account or external PDF reader is needed for this course explanation.

Read the short guide here

Four habits for working with AI. This original course explanation applies the AI Fluency Framework to your project; it is not a reproduction of the publisher's one-page guide.

Delegation — choose the work
Decide whether AI is useful for this task, which step it may assist, and which decisions a person must retain.
Description — make the task clear
Give the permitted inputs, goal, audience, output format, constraints, and a clear stopping rule.
Discernment — inspect the result
Compare the output with sources and success criteria. Check for missing conditions, invented claims, and whether it helps complete the intended task.
Diligence — remain responsible
Own the decision to use or release the result. Protect information, disclose material AI assistance, and arrange human review or escalation where needed.

Apply to your task: Write one AI-assisted step, one instruction, one output check, and one responsibility you keep. Use your existing project record; this is optional practice, not another submission.

Framework attribution: Rick Dakan, Joseph Feller, and Anthropic, The AI Fluency Framework (2025). The definitions above are course-authored paraphrases and project prompts. The original framework guide is released under CC BY-NC-SA 4.0.

Official framework — original one-page PDF ↗

Try

Try / predict

Handle a source change without guessing

The owner says the cancellation policy has changed but has not supplied approved wording. An old draft is ready to send. Predict the next action. A plausible new policy generated by AI is not an authorized replacement; pause the affected output and obtain the approved version.

Worked recovery sequence

Mark the old output held. Record which source and cases may be affected. Obtain and label the new source, update the response rule, and rerun boundary and exception tests. Ask a reviewer to approve the revised version before use. Preserve a brief change log so users know which version is current.

If an incorrect reply had already been delivered in an actual service, the owner would assess the impact and authorize an appropriate correction. In this classroom scenario, describe that procedure; do not invent a sent correction or customer notification.

Now give your draft instructions to someone unfamiliar with the kit. Ask them to locate the current source, identify who can authorize an exception, and find the fallback. Difficulty finding these items is a handoff defect, even if your demonstration was smooth.

Decision lab · practice with your Learning Companion

The queue at closing time

Plan for review capacity, exceptions, and a stopping rule.

Take the role of operations owner. Commit to a decision, uncover evidence, revise your reasoning, then carry one action into Build. Use this as the unit’s reasoning rehearsal; reuse the response in your existing project update.

Open the AI Learning Companion section below this interactive lesson to enter this unit’s decision lab. Return here for Build. You can also use the complete reading route below.

Complete reading route · no AI service needed

Fictional teaching scenario. A separate fictional workflow receives 18 drafts per day, but its owner can review only 12 per day. All consequential replies require review. Choose the operating response.

  1. Auto-send the excess six drafts
  2. Cap or pause intake and route urgent exceptions to an owner
  3. Keep accepting work without a warning

Choose one and write why before revealing the evidence.

Reveal evidence and challenge
Supplied evidence
  • Queue arithmetic: 18 arrivals − 12 reviews = 6 additional unreviewed drafts per day while those rates persist.
  • Capacity limit: After three such days, 18 additional drafts are waiting, assuming an initially empty queue and no other changes.
  • Source change: The approved source version changes during the backlog; pending drafts need a source check before release.

Write a stop rule and recovery step. Who owns the queue, the source update, and communication about delays?

Compare with the worked reasoning
Worked reasoning

At unchanged rates the backlog grows by six per day. Cap intake or add approved review capacity; route urgent cases and communicate delays. After a source change, check pending drafts against the new approved version before sending. Name an owner, a queue limit, a fallback, and a restart check. AI drafting speed does not create human review capacity.

Apply to your project: Write your project’s review capacity, one pause trigger, its owner, and the evidence needed to restart.

Write your initial choice, evidence-based revision, and project action in your own document.

Practice boundary

Practice records are not grades or Canvas submissions.

Build

Apply this step to your chosen problem. Cedar Desk is the supplied classroom case; keep one project throughout the course.

Complete a short operating pack

Write a Start page, a one-page operating checklist, and a change/issue log. Link to existing materials instead of copying conflicting versions. Add the final source version, reviewer, owner, support limit, and continue/revise/stop rule.

Try the handoff on an ordinary and changed-information scenario. Record the actual attempt and improve the instructions. For a partner project, agree on ownership, permitted continued use, and what happens to access after the course. For a classroom project, state that it remains a demonstration using fictional inputs unless a later approved arrangement changes that status.

Describe how you would learn about repeat use or abandonment, with a named cohort and follow-up period. The supplied four-of-six return figure from Unit 7 is a practice example, not a retention forecast for your service. The next unit brings the evidence and usable product together in a decision-focused demonstration.

Add your project evidence and reflection to the check-in in Canvas.

Your project record

Write here or use your own document. Notes stay in this browser; export a backup to keep them. Nothing is submitted automatically.

Finish

What will you carry forward?

Your Canvas check-in

The policy may have changed, but the approved replacement is unavailable. What is the sound operating response?

Use the evidence from this lesson to explain your reasoning in Canvas. Continue with Canvas Next below; submit your check-in in Canvas.

The simulation and Learning Companion support practice. They do not submit work, record Canvas completion, or assign a grade.

The policy may have changed, but the approved replacement is unavailable. What is the sound operating response?

In your project record, state what you decided, the evidence you used, and one uncertainty to test next. Explain the reasoning in your own words.

You can revisit any step. Completion records study progress; it does not confirm a submission, grade, or project quality.