1. The "sampling log" self-monitoring pattern involves:
Correct. The sampling log is low-tech but powerful: save random outputs, read them regularly. Gradual drift is hard to hide from a human who reads real outputs rather than just checking dashboards.
The sampling log pattern saves a random sample of real outputs for regular human review. It's low-tech but catches the gradual drift that dashboards and error logs miss.
2. You're evaluating two workflow tool choices: Zapier vs. n8n. A nonprofit wants to build a workflow that processes donor information and must not store that data on any external server due to privacy commitments. Which tool is the better fit and why?
Correct. The lesson specifically describes n8n as the tool for situations where sensitive data must not go through third-party cloud servers. It runs locally, giving the nonprofit full data control.
n8n's distinguishing feature from the lesson is exactly this: it's open-source and can run on your own infrastructure. Zapier stores workflow data through its cloud. For a privacy requirement that prohibits external storage, n8n is the correct choice — not because it's simpler (it isn't), but because of data sovereignty.
3. You're presenting an automation to your school's IT committee. The most effective opening, according to this module, would be:
That's the before-and-after frame in action. Specific current cost (2 hours, 40–60 tickets, manual, Monday) followed by a specific change (automatic, at submission). The gap is concrete and immediate.
Abstract claims about transformation, competitive comparisons, and technical details all fail the before-and-after test. Only option B names a specific current cost and a specific change — the gap the audience can feel.
4. A workflow runs automatically every night to back up files. One night, the cloud storage service is temporarily offline. The workflow produces no error and logs "backup complete" — but nothing was actually saved. Which failure pattern is this?
Yes. A wrong result that looks right, with no visible error — that's a silent failure. The log says "complete" but nothing was saved. This could go undetected indefinitely.
Silent failure is the pattern where the workflow produces a wrong output that looks correct, with no error message. "Backup complete" logged when nothing was saved is the textbook version of this.
5. A school uses an AI workflow to automatically flag students' social media posts as "concerning" and notify their parents. The optimization target is nominally "student wellbeing." What hidden value choice is most embedded in this system?
Exactly. "Concerning" is not a neutral category — it embeds someone's values about what speech is acceptable. The hidden choice is: whose definition of concerning, and who decided that standard applies to students' private expression?
The deepest hidden value is in the definition of "concerning" itself — that label requires someone to decide what kinds of expression are problematic, and that decision encodes cultural and political values that students never agreed to be governed by.
6. Clearview AI's database of three billion images was built primarily using data scraped from:
Correct. The images came from public social media — a use that the people posting those images could not have anticipated when they shared them.
Clearview scraped public social media — photos people had posted publicly, but without any expectation those images would end up in a commercial biometric database sold to law enforcement.
7. Steven Schwartz was sanctioned because he submitted AI-generated fake legal cases to a court. What was the most fundamental design error in how he used the AI?
Correct. The mismatch was between what the input assumed (that the AI can retrieve real legal records) and what the AI actually does (generate fluent text). Input design must match the tool's real capabilities, not assumed ones.
The problem wasn't query detail or the pricing tier. It was a fundamental mismatch: the input assumed the AI could verify and retrieve real court records, when in fact it generates statistically likely-sounding text. That's an input-design failure.
8. Which of the following best describes "responsibility diffusion" in AI systems?
Correct. Responsibility diffusion is the accountability gap that emerges when harm crosses organizational boundaries within a pipeline. Each party can point to the next — and the person harmed may have no clear legal recourse.
Responsibility diffusion is specifically about accountability becoming unclear when an error spans multiple parties' work. It's not a technical architecture — it's a social and legal phenomenon that emerges from multi-party pipeline design.
9. Why do people inside broken processes often stop seeing the broken seams?
Yes. Adaptation makes broken things feel normal. That's why an outside perspective — someone who hasn't adapted — can see the seam that insiders have long since stopped noticing.
It's not about technical knowledge or policy — it's about adaptation. When you live inside a broken process, you build workarounds and the friction stops feeling like friction.
10. IBM's Watson Health project failed primarily because:
Exactly. The scope trap. A challenge that's too broad to define clearly is too broad to solve.
The technology wasn't the problem — the definition was. "Help treat cancer" can't be solved by any tool because it's not specific, bounded, or testable.
11. A "silent failure" in an AI workflow is best defined as:
Correct. Silent failures are the hardest to catch because everything appears to work — the error hides in the output, not in the execution log.
Silent failures are defined by their invisibility: the workflow runs, reports success, and produces output — the output is just wrong. No error signal, no crash.
12. An "upstream error" in a workflow means:
Correct. Upstream errors are dangerous because every downstream step processes bad data without knowing it — each step reports success, but the error compounds through the chain.
Upstream errors start early and travel silently. Every step after the broken one processes the corrupted data faithfully and reports success — making the source hard to locate.
13. Informed consent requires three things. Which of the following is NOT one of them?
Correct. The right to deletion is a separate legal right in some jurisdictions — but it's not one of the three components of informed consent as defined in the lesson (told clearly, understand it, genuine option to decline).
The three components of informed consent are: being told clearly, understanding what is happening, and having a real choice to say no. The right to deletion is a separate legal concept, not one of the three.
14. What is the most important habit to develop when building multi-step AI workflows, based on Lesson 2?
Right. Intermediate output inspection is the core debugging practice. Errors in the middle of a chain are invisible from the end — you have to look at what's passing between steps.
The lesson emphasized inspecting intermediate outputs specifically. Looking only at the final output — after all steps have run — means a problem in Step 2 has already influenced Steps 3, 4, and 5 before you catch it.
15. A workflow auditor checks a 5-step automation. Steps 5, 4, and 3 all show correct inputs and outputs in the execution log. Step 2's output looks unusual — it contains only a customer's email subject line, not the full email body. Step 1 is the trigger (an incoming email). Where is the root cause?
Exactly right. Step 2's output is the first unusual output in the trace-back. The root cause is either in step 2's configuration (field mapping to the subject line instead of the body) or in how step 1 passes data to step 2.
The trace-back found step 2's output as the first anomaly. The root cause is either in step 2 itself (wrong field mapping) or in how step 1 passes data to it. That's where good input (full email) became bad output (subject line only).
16. What was missing from Facebook's content moderation workflow in September 2016, according to this module?
Correct. The workflow could act but couldn't hesitate. A confidence check or uncertainty-flagging mechanism would have routed the unusual photograph case to a human reviewer.
The missing component was a self-monitoring layer — a way for the system to recognize unusual cases and route them to human review before taking irreversible action like content removal.
17. Bias laundering in AI systems is most dangerous because:
Correct. The danger is the disguise: bias that arrives wearing math looks neutral, which is why people accept it without the scrutiny they'd apply to a human making the same biased call.
Bias laundering is dangerous because it makes unfairness look like objectivity — dressing inherited historical bias in the authority of statistics.
18. The "accountability gap" in automated systems refers to:
Correct. The Elaine Herzberg case illustrated this clearly: the harm came from a distributed system, and no single actor could be clearly identified as "the one who decided."
The accountability gap is specifically about responsibility distribution: when decisions are made by systems involving many actors, it becomes hard to say who is legally and morally responsible for the outcome.
19. "Prompt rot" occurs when:
Correct. Prompt rot is slow and invisible — the prompt is unchanged, but the world around it (model, data, context) has shifted. Output quality degrades gradually without any error signal.
Prompt rot is gradual, invisible, and doesn't require anyone to change the prompt. It happens because the model or the data around the unchanged prompt has shifted.
20. The "stakes calibration rule" for deciding when to keep a human in the loop focuses on:
Right. Stakes calibration is about consequence: high-stakes, irreversible, person-affecting decisions warrant human oversight. Low-stakes, easily reversible decisions can be fully automated.
Stakes calibration focuses on the consequences of errors — their cost, reversibility, and whether they affect real people — not on technical parameters like model age or step count.