1. The IBM Watson Health failure at MD Anderson Cancer Center cost approximately:
Correct. IBM's total losses on Watson Health exceeded $4 billion before the assets were sold or shuttered in 2022.
IBM's total Watson Health losses exceeded $4 billion — making it one of the most expensive change management failures in AI history.
2. Which of the following is a LAGGING indicator of AI adoption (as opposed to a leading indicator)?
Correct. Annual revenue per employee is a lagging indicator — it confirms adoption occurred but arrives 6–12 months too late to allow timely course correction.
Annual revenue per employee is a lagging indicator — it confirms results after the fact. Manager adoption rate, confidence scores, and peer knowledge sharing are leading indicators that appear early enough to drive intervention.
3. What happened to Adidas's German Speedfactory, and why is it instructive for understanding automation?
Correct.
Incorrect. Adidas closed the Speedfactory in 2019 due to insufficient flexibility — a real-world lesson that full automation optimizes for volume and consistency, not variety.
4. What did the 2023 WGA agreement say about writers who use AI tools with studio approval?
Correct.
The WGA agreement protected writers from having their credit or guaranteed compensation reduced because they used AI tools the company approved — closing a potential loophole.
5. What is a "cobot"?
Correct.
Incorrect. Cobots are collaborative robots designed to work alongside humans — increasingly used in small and medium manufacturing.
6. What does Denmark spend on active labor market policies as a share of GDP, compared to the United States?
Correct. The OECD's 2023 Employment Outlook noted Denmark spends ~2% of GDP on ALMPs versus ~0.1% in the U.S. — a 20x difference reflecting fundamentally different policy philosophies.
According to OECD data, Denmark spends approximately 2% of GDP on active labor market policies; the U.S. spends approximately 0.1% — a roughly 20x difference.
7. GitHub Copilot operating inline in VS Code is the canonical example of which model?
Correct. Copilot's inline suggestions — always present, no context switch — define the cyborg model.
Not quite. Copilot is cyborg: continuous AI presence embedded in the workflow, no discrete handoff moments.
8. The Acemoglu-Restrepo (2018) "Race Between Man and Machine" framework centers on the balance between which two forces?
Correct.
Acemoglu and Restrepo's framework identifies displacement (automation reduces human labor demand in automated tasks) and reinstatement (new tasks create new human labor demand) as the competing forces determining net employment outcomes.
9. What structural mechanism did both the WGA and SAG-AFTRA 2023 agreements include to address future AI developments?
Correct.
Both agreements created joint labor-management AI committees to govern new AI applications as they emerge — acknowledging the need for ongoing governance rather than fixed rules.
10. Prosci's research finds that organizations actively managing all five ADKAR elements achieve what advantage over those focused only on awareness and training?
Correct. Prosci's 2023 benchmarking found 6× higher project success for organizations managing all five ADKAR elements versus those focused only on awareness and training.
Prosci's 2023 benchmarking found 6× higher project success for organizations managing all five ADKAR elements.
11. The Humlum Danish firm study found that higher wage pass-through from AI gains was associated with:
Correct. Institutional worker representation — works councils and collective agreements — was the key variable associated with higher wage pass-through from AI productivity gains.
The Humlum study found that works councils and collective agreements predicted higher wage pass-through — institutional bargaining power shapes distribution of AI gains.
12. Under the EU AI Act, employment AI systems used for performance evaluation must be registered where?
Correct.
High-risk AI systems used in employment must be registered in the European AI Office's public database — a transparency mechanism allowing workers, unions, and regulators to see what systems are in use.
13. "Task exposure," as used in the Eloundou et al. (2023) paper, most precisely means:
Correct.
Task exposure is the degree to which AI can perform a task at equivalent or lower cost — distinct from job elimination. High exposure doesn't mean the job disappears; it means that task may shift to AI within a restructured bundle.
14. What did the 2023 WGA agreement establish about AI-generated material as "source material"?
Correct.
Incorrect. The WGA agreement bars AI output from being "source material" — closing the loophole of paying writers reduced rates to merely adapt AI-generated work.
15. What did the Brynjolfsson 2023 Stanford study find about workers who stopped engaging critically with AI outputs?
Correct. Workers who accepted AI outputs without review showed quality regression — critical engagement is the mechanism that produces both productivity and quality gains.
Not quite. Passive AI acceptance led to quality regression. Critical engagement — not just AI access — is the mechanism of augmentation.
16. Microsoft's 2024 Copilot adoption study identified a "30-day cliff" where employees reverted to prior workflows unless they had received:
Correct. Employees cleared the 30-day cliff when they had received structured guided application sessions AND seen clear productivity evidence from their own use — both were necessary.
Microsoft found that structured guided application sessions combined with personal productivity evidence were the two factors that prevented regression past the 30-day cliff.
17. Mass General Brigham radiologists using AI lung nodule detection experienced which outcome?
Correct. Mass General Brigham radiologists gained ~30% speed and improved clinically significant finding rates — a textbook augmentation outcome with the AI handling the mechanical pass and humans retaining judgment.
Not quite. Mass General Brigham radiologists working with AI lung nodule detection saw ~30% speed gains and improved significant finding rates — augmentation at work.
18. Deloitte's 2023 survey found that replacement strategies apply as genuinely preferable to reskilling in approximately what percentage of roles undergoing AI transformation?
Correct. Deloitte found that replacement is genuinely preferable (due to fundamental role change) in only ~12% of roles — reskilling is more cost-effective for the vast majority.
Deloitte found replacement strategies preferable to reskilling in only ~12% of roles — where roles change so fundamentally that adjacent skill pathways genuinely don't exist.
19. In the MIT Noy and Zhang study, workers with ChatGPT completed writing tasks in an average of how many minutes?
Correct. The MIT study found 17 minutes with AI versus 27 minutes in the control group — a substantial time reduction.
The MIT study found 17 minutes with AI versus 27 without, representing a roughly 37% time reduction.
20. MLOps extends DevOps principles to ML systems primarily to:
Correct. MLOps applies software engineering rigor to ML systems — standardizing the path from model development through production deployment, monitoring, and retirement.
Incorrect. MLOps applies DevOps principles (continuous integration, delivery, and monitoring) to ML systems — standardizing and automating the entire model lifecycle to enable reliable production deployment at scale.