Kuis AI & Pekerjaan

Jawaban dan penjelasan muncul saat Anda mengikuti kuis.

1. An occupation has a high AI exposure score. What does that score establish?
   • Potential task overlap, not a direct forecast of job loss
   • That every worker will be replaced
   • How many new jobs will be created
   • The date a workplace will adopt AI
2. Why analyze tasks within a job rather than treating the occupation as one unit?
   • Every task in an occupation has the same technical requirements
   • A role combines tasks that may be automated, augmented, or still need human context
   • Occupation titles fully describe daily work
   • One automated task means the role can be removed
3. A model can draft a report. Why might an organization still not adopt it?
   • Technical feasibility guarantees deployment
   • Only model size determines adoption
   • Reliability, cost, integration, regulation, and work choices affect adoption
   • Worker training is irrelevant
4. A tool speeds up one stage but adds a large review queue. What should be evaluated?
   • Only generation speed
   • The number of prompts used
   • The model’s benchmark score alone
   • Total work and job quality after review and correction are included
5. Which statement distinguishes augmentation from replacement?
   • A system may assist one workflow while automating part of another
   • Every AI use either fully replaces or never changes a job
   • Augmentation means no work changes
   • Replacement can be inferred from task exposure alone
6. What labor-market evidence should inform a career decision?
   • A global exposure estimate used as a personal forecast
   • Current information for the specific location or occupation
   • A single company’s product launch
   • An old job title list without a date
7. Which capability can help workers evaluate AI-supported tasks?
   • Prompt memorization without subject knowledge
   • Avoiding all review responsibility
   • Domain knowledge, output evaluation, exception handling, and clear communication
   • Choosing a model by parameter count
8. Why should workers be able to challenge inaccurate AI outputs?
   • A challenge process guarantees no errors
   • Only managers can detect mistakes
   • A model recommendation removes human accountability
   • They may carry error responsibility and need a way to correct outcomes