Session 1 · Fair and accountable AI: ethics and bias · Lesson 10 of 29
Session 1 activities
ACTIVITIES · 33 MINFictional data only
All companies, people, and data are fictional. Never enter real customer, patient, or employee information (G.L. c. 93H, 201 CMR 17.00).
You can do every activity here. Sign in to save your responses and add your vote to the class tally.
Who is accountable?
Poll at the start of session 1, then discuss.
North Shore Builders' (fictional) AI scheduler never offered weekend overtime to one crew. Who is accountable?
Bias detective
Match each fictional example to the type of bias it shows most clearly (Svekis, chapter 9), then name who you would tell.
Buckets: Historical biasRepresentation biasMeasurement biasEvaluation bias
- Resume model trained on ten years of mostly male hires ranks women lower
- Voice tool misunderstands strong regional accents
- Scheduler treats 'hours badged in' as effort, so workers who log site visits by phone look less productive
- Safety-glasses detector was tested only on the day shift under bright lights
- Image tool shows only men when asked for 'a construction manager'
- Credit model scores customers from a rural county poorly because few rural loans were in its data
- Call-quality model rates calls by length, so short, efficient calls score as 'poor service'
- Translation tool approved after testing only on English and Spanish, then used for Portuguese-speaking customers
0 of 8 sorted
The scheduler
Branching. You notice the North Shore Builders scheduler pattern described in Lab 1.
What do you do first?
Ethical risk check
After session 1, post.
Choose one AI use you could see at work. Using the five principles, name the main risk and one question you would ask before using it.
- The use.
- The principle most at stake.
- Who could be affected.
- One question to ask.