When six ethics lenses disagree on a real AI decision, which one do you let win — and what does that cost?

6-Lens Case Analysis

Context

You are an editorial desk commissioned to produce a special issue, 'Machines That Decide.' You have one real AI case to dismantle properly — not a hot take or a vibe, but disciplined reasoning through all six Markkula ethics lenses. The system harmed real people. Six different ethical frameworks point in different directions. Your editor will not print hedging.

Mission

Produce an ~800-word 6-Lens Case Analysis running one real AI case (Rotterdam, COMPAS, Dutch childcare-benefits scandal, exam-grading system, or hiring filter) through the Markkula framework — rights, justice, utility, common good, virtue, care — isolating at least two lenses that contradict each other on this case, and ending with a defensible verdict that names which lens wins, what the choice costs the other stakeholders, and what accountability remains for the harmed.

Finish Line

An ~800-word 6-Lens Case Analysis ending in a defensible verdict, fit to print.

Deliverables

6-Lens Case Analysis
lesson

A written analysis of one AI case run through all six Markkula lenses.

Team Roles

Case Lead

Owns the facts and the frame
- Produces a 2–3 sentence opening that names the system, its decision, and the group harmed — precise numbers, no conjecture, no value-judgment framing. (Example: 'In 2009, the city of Rotterdam used an algorithm to flag benefit applicants as fraud risks. The algorithm flagged 30,000 single mothers. The city acted on the flags; a court later ruled the system illegal.')
- Challenges every lens argument that rests on a guessed fact or softened consequence — pushes back on Lens Counsel and Verdict Editor when they invent details or hedge the real harm.
- Locks the case to a verifiable source (court ruling, news archive, published audit, investigative report) — team cannot proceed until Case Lead confirms the facts can be cited in print.
Lens Counsel

Steelmans all six, especially the one you hate
- Argues each lens at full strength — produces a distinct paragraph (200+ words) for each of the six frameworks, giving the strongest version even when it points away from personal view. Tests: would someone who centers that lens as their primary frame accept it without calling it a strawman?
- Isolates real tension: identifies where Rights, Justice, Utility, Common Good, Virtue, and Care differ in their core claims, not just their conclusions. (Rights is not Justice; Utility is not Care; confusing them collapses the analysis.)
- Commissions rewrites of any lens reduced to a caricature — if a paragraph could apply to three different lenses, it is not yet written.
Verdict Editor

Forces collision, names cost, holds the ruling
- Isolates at least two lenses that point in opposite directions on this specific case and explains why the conflict is structural, not word-play. Extracts from the final text: lens A says X, lens B says Y, they cannot both be true.
- Extracts from the final verdict: which lens wins, what values and stakeholders lose under that ruling, and a claim addressed directly to the harmed party. Rejects any verdict missing a named loss.
- Enforces clear, unambiguous judgment — cuts hedging ('it depends,' 'there are good points on both sides,' 'one could argue'). The ruling must survive a hostile editor's line-by-line challenge in print.

Exemplars

Anki — powerful, intelligent flashcards
AnkiWeb

Gold-standard personal mastery system: active recall + spaced repetition. Validity proven by repeated solo testing, not by looking things up — the capstone’s whole logic.