The one-sentence truth: AI becomes a useful GP topic when we stop treating it as one machine and start examining capabilities, incentives, tasks, institutions and accountability.
Core vocabulary
Artificial intelligence, generative AI, automation, augmentation, algorithmic bias, accountability, human oversight, alignment, scalability, verification.
Capability versus autonomy
A system can perform complex tasks without possessing human-like agency. Be precise about what capability is being discussed rather than anthropomorphising technology.
Automation and augmentation
AI may substitute for tasks, complement workers or create new workflows. Employment arguments should therefore be task-specific.
Bias
Algorithmic bias can arise from data, objectives, measurement choices or deployment context. “The algorithm is biased” is only the beginning of analysis.
Accountability
When AI informs decisions, responsibility still belongs somewhere: developers, deployers, institutions or professionals. Automation can redistribute responsibility without eliminating it.
Verification
Generative systems can produce plausible outputs that require checking. Their value depends partly on the user’s ability and incentive to verify.
Example sentence
“Generative AI is most likely to augment human judgement where outputs can be independently verified; in high-stakes settings, speed without accountability can amplify rather than reduce risk.”
Singapore application
Use education, work, public services and digital policy as contexts while distinguishing productivity gains from governance, distribution and trust.
Continue the JC2 GP vocabulary system
Return to General Paper Vocabulary Year 2 (JC2) and Vocabulary for Junior College (JC1–JC2).
Final thought
The central AI question is not whether machines become intelligent in the abstract, but how new capabilities change human decisions, incentives and responsibility.