Responsible AI in Practice · Section B · 46 learners

Tutor settings by activity

Program policy: Graduate certificates. Open mode isn't allowed on graded work. Transcripts are kept for 1 year. You can choose anything inside those limits. View policy · set by Academic Technology

Module 3 activityTutor mode

3.1 Why evaluation misleadsLesson · practice OffHintsExplainOpen

3.2 Bias in model evaluationLesson · practice · includes a quick check OffHintsExplainOpen

Scenario: Defend the metricAI role-play · practice · ungraded OffHintsExplainOpen

Module 3 checkGraded quiz · 3 questions · editing, changed from Explain OffHintsExplainOpen

Essay 2: Fairness audit memoGraded assignment · AI use must be declared OffHintsExplainOpen

Off: No tutor on this activity.
Hints: Questions and hints only. Never gives answers.
Explain: Explains concepts and examples, not item answers.
Open: Answers directly. Practice only.

Module 3 check · Hints

How the tutor behaves here

Hints per question

Worked examples Parallel cases only (same idea, different numbers; recommended for graded work) No worked examples

When a learner asks for the answer Decline, give the next hint, and count it in the summary you see Decline and suggest office hours

Sources the tutor may use Week 3 slides Reading 3.2 and 3.3 Fairness audit guide Lecture 3 recording (transcript) Answer key · never available to the tutor

What you'll see: a per-learner summary (hints used, answer requests, topics), plus class patterns in Insights. Full transcripts open only from a submission, and learners are told that.

Preview as a learner

Not saved yet

Notice learners see: "Hint mode, set by Dr. Okafor for this check. The tutor answers only from course materials, and your instructor sees a summary of this chat."

Which option is right for question 2?

Course tutor · hint 1 of 2

I can't give answers on this check, but here's a way in: if you delete the ID column, can the model still recognize the same patient from their other features?¹

Week 3 slides, p. 6

Just tell me.

Course tutor · hint 2 of 2

The fix has to change which rows land in each set, not which columns the model sees. That's the last hint for this question.