AI at work · Case analysis · 50 min
What goes in the tool, and who gets told
The two questions that get entry-level workers in trouble: what data was I allowed to paste, and did I say a tool was involved.
Target level: Proficient — Has functional knowledge of the skill and can apply it properly in specific on-the-job settings.
Why this design
Work a real situation from evidence to recommendation, and show the reasoning. Judgment content — root cause, tradeoffs, competing priorities, risk.
What this evidences
- data-privacy-judgment×1 at Proficient
- ai-ethics-disclosure×0.6 at Proficient
By the end, the learner can
- Automated network audit pipeline: orchestrates AI tools to scan running configurations against compliance benchmarks (CIS, NIST), triage flagged deviations by risk severity, and produce remediation tickets, reviewing every high-severity finding personally.
- Interpretation gate-keeping: evaluates AI-generated choropleth or heat-map outputs against domain knowledge of local land use and demographics, overriding or correcting automated classifications that misrepresent on-the-ground conditions.
- Access control change validation: uses AI tooling to cross-reference proposed individual access-status modifications against role-based policy rules, catching errors before committing changes to security files.
Session plan
- 15 minThe caseWhere this shows up: technology teams and corporate it environments, working with managers, technical teams, users. Conditions to expect: analytical, system-dependent, problem-solving intensive. Entry roles this maps to: Computer Systems Analysts, Computer Systems Engineers/Architects, Computer User Support Specialists.
- evaluates AI-generated choropleth or heat-map outputs against domain knowledge of local land use and demographics, overriding or correcting automated classifications that misrepresent on-the-ground conditions.
- 15 minEvidence logWork the task on your own first. Support is available, but ask for it out loud so the group hears what got you stuck.
- Tools in play: Development environment software, Data base user interface and query software, Web platform development software.
- 10 minCompare readingsCompare answers before the debrief. Where the room split is the useful part — the split usually marks a real judgment call rather than a knowledge gap.
- Where did two people read the same situation differently, and what did each notice?
- Which answer would survive a supervisor asking "why did you do it that way?"
- 10 minWhat the record showedClose the loop: what changed between the first attempt and the last, and what would have to be true for data & privacy judgment to hold up on a bad day.
Assessment (17 items)
Loading your attempt…