Technology and AI Vocabulary for JC General Paper

Technology and AI Concept Lab: Build the Argument Behind the Buzzwords

Technology vocabulary becomes valuable when it separates different kinds of change. “AI”, “automation” and “innovation” are too broad to carry an argument on their own. Students need concepts that describe what the technology does, who gains capability, who loses control and how institutions respond.

Core AI Distinctions

• Automation: transferring a task from human execution to a system.
• Augmentation: increasing human capability while keeping the person in the task.
• Autonomy: degree to which a system can act without immediate human direction.
• Assistance: providing recommendations or drafts while a person remains responsible for the decision.
• Substitution: replacing human labour or judgement in a particular function.

An essay about “AI replacing people” becomes much stronger when it asks which tasks are automated, which are augmented and which still require human accountability.

Bias, Fairness and Representation

Bias can enter through training data, measurement, objectives, deployment context or human interpretation. Fairness is not one universal standard: equal treatment, equal error rates, equitable outcomes and procedural fairness can point in different directions. Students should avoid writing “the algorithm is biased” without identifying the mechanism and affected group.

Transparency, Explainability and Interpretability

These terms overlap but are not identical. Transparency concerns visibility into processes or rules. Explainability concerns whether reasons for an output can be communicated meaningfully. Interpretability often concerns how understandable the system’s internal logic or relationships are. A high-stakes system may require more than a simple output; affected people may need reasons they can challenge.

Accountability and Human Oversight

If a system recommends a decision, who remains answerable when it causes harm? Useful terms include human oversight, appeal mechanism, auditability, traceability and responsibility gap. The key distinction is between using a tool and outsourcing accountability to it.

Labour Vocabulary

• Displacement: existing tasks or roles are reduced or removed.
• Task reconfiguration: jobs remain but their internal mix of tasks changes.
• Deskilling: repeated reliance on automation may reduce opportunities to practise a skill.
• Reskilling: workers develop capabilities for new roles or task mixes.
• Productivity gain: greater useful output from the same or fewer inputs.
• Transition cost: short-term losses or disruption during adaptation.

Power and Concentration

Technology can democratise access while simultaneously concentrating infrastructure, data or market power. Vocabulary such as network effect, market concentration, platform dependence, interoperability and switching cost helps students explain why a useful service can also create structural dependence.

Privacy and Surveillance

Personalisation often depends on data collection. The tension is not simply privacy versus convenience. Ask what data are collected, whether consent is meaningful, how long data persist, whether new uses are compatible with the original purpose, and whether individuals can realistically opt out.

Innovation Versus Precaution

Precaution does not mean stopping every uncertain technology. It means adjusting the burden of proof when potential harm is serious or difficult to reverse. Regulatory sandbox, pilot, staged deployment and risk-tiered regulation are useful concepts because they create options between unrestricted deployment and prohibition.

Worked GP Example: AI in Education

A strong argument might distinguish routine assistance from cognitive substitution: “Generative tools can improve access to feedback and examples, but their educational value depends on whether they augment student thinking or replace the retrieval, planning and revision processes the student is meant to learn.” This opens further questions about assessment validity, authorship, teacher workload and unequal access.

Singapore Transfer Questions

• Where can digital scale improve access to public or educational services?
• Which high-stakes uses require explicit human review?
• How should institutions preserve accountability when vendors provide the underlying systems?
• What skills become more valuable when routine cognitive work is automated?
• How should access, privacy and innovation be balanced in a highly connected society?

Technology Vocabulary Drill

Choose one technology and ban yourself from using the words good, bad, useful and dangerous. Instead describe its effect through five concepts: capability, incentive, risk, accountability and distribution. Then write a judgement that identifies the condition under which your position changes.

Technology essays become sophisticated when the vocabulary moves the discussion from devices to systems: what is automated, what is amplified, who is accountable, who bears risk, and which human capabilities society wants to preserve.