Technology discussions become weak when every tool is described as “good,” “bad,” “smart” or “dangerous.” Students need vocabulary that separates what a system does, what evidence supports its benefits, what risks exist and who is affected.
This article complements the Secondary 2 Essential 100 Vocabulary by extending words such as analyse, bias, evaluate, innovation, manipulate, validate and significant into contemporary technology contexts.
Core Technology Vocabulary
| Word | Working meaning |
|---|---|
| algorithm | a defined procedure or set of rules for processing information or solving a problem |
| automation | using systems to perform tasks with reduced direct human intervention |
| artificial intelligence | computer systems designed to perform tasks associated with capabilities such as prediction, recognition, generation or decision support |
| data | recorded information used for analysis or processing |
| model | a representation used to describe, predict or generate outputs |
| bias | a systematic tendency that can distort outcomes or judgement |
| privacy | control and protection relating to personal information |
| verification | checking whether information or a result is accurate or valid |
Do Not Treat AI as One Thing
An AI system used to recommend videos, recognise images, generate text or detect anomalies performs different jobs. Good vocabulary begins by naming the task before judging the technology.
Vocabulary for Benefits
- efficient;
- scalable;
- adaptive;
- accessible;
- beneficial;
- innovative;
- resourceful;
- automated.
Each benefit needs evidence. Calling a system efficient requires some account of time, cost or resources saved.
Vocabulary for Risks
- biased;
- inaccurate;
- opaque;
- manipulative;
- misleading;
- invasive;
- unverified;
- vulnerable.
Again, avoid automatic labels. Ask what evidence justifies the risk description.
Technology and Responsibility
Useful vocabulary includes accountability, consent, transparency, legitimate use, safeguard, oversight and responsibility. These words help students discuss not only what technology can do, but who should decide, verify and respond when something goes wrong.
AI and Schoolwork
Students can ask whether a tool enhances learning or merely substitutes for thinking. Does it help a learner interpret, analyse or practise? Can the student validate the output independently? Can the student acknowledge where assistance was used?
A Discussion Model
“Generative AI can be beneficial when it helps students explore explanations or receive feedback, but its value depends on verification and independent understanding. If students accept outputs without evaluation, errors or bias may be incorporated into their work.”
Avoid Hype Vocabulary
Words such as revolutionary, transformative, terrifying and inevitable should not replace analysis. Ask what has actually changed, for whom and under what conditions.
A Technology Vocabulary Practice Routine
- Choose one technology.
- Name its actual task.
- Identify one benefit.
- Identify one risk.
- State what evidence would validate each claim.
- Identify one responsibility or safeguard.
Technology Vocabulary Should Increase Resolution
Use the Secondary 2 Essential 100 as the reasoning base, then add technology-specific language. The goal is not to sound futuristic. It is to discuss systems precisely enough to distinguish capability, evidence, benefit, risk and responsibility.
AI Literacy Laboratory: Separate Task, Capability, Evidence, Risk and Responsibility
The most useful way to discuss artificial intelligence is to resist treating “AI” as a single agent with one ability. Start with the task. What input enters the system? What output is produced? What capability does the task require? What evidence shows that the output is reliable enough for this use? What happens if it is wrong?
The Five-Layer AI Question
- Task: What is the system being asked to do?
- Capability: What can it do well enough to be useful?
- Evidence: How do we know?
- Risk: What failure matters in this context?
- Responsibility: Who must verify, decide or intervene?
This structure remains useful even as particular products change. It teaches durable technology vocabulary rather than dependence on today’s brand names.
Worked Example: AI Feedback on a Student Paragraph
Task: identify unclear sentences and suggest improvements.
Possible benefit: the system may provide rapid feedback and expose the student to alternative phrasing.
Risk: a suggestion may be inaccurate, may misunderstand the intended meaning, or may make the language more elaborate without making it better.
Verification: the student compares the suggestion with the original purpose, checks factual claims independently, and decides whether the revision improves precision.
Responsibility: the learner remains responsible for understanding and defending the final writing.
Generation Is Not Verification
A system can produce a fluent answer without independently establishing that every statement is correct. This is an important vocabulary distinction: generate, predict, retrieve, verify and validate describe different jobs. Fluency should not be confused with evidence.
Automation Bias
People may give extra weight to a system’s output simply because it was produced automatically or presented confidently. This tendency is often described as automation bias. The practical lesson for students is simple: a polished interface does not remove the need to evaluate the underlying claim.
Human Oversight Has Different Levels
| Role | Human involvement | Example |
|---|---|---|
| tool | human chooses and evaluates each use | suggesting alternative sentence wording |
| assistant | system proposes; human reviews before action | drafting a study plan |
| automation | system performs routine task with monitoring | sorting large sets of low-risk items |
| high-stakes decision support | requires strong evidence, safeguards and accountable human judgement | contexts where errors could seriously affect people |
Bias Needs a Mechanism
Calling a system “biased” should lead to a second question: biased how? Possible mechanisms include unrepresentative training data, unequal error rates, labels reflecting earlier human decisions, or a design objective that rewards one outcome over another. Precision prevents bias from becoming a vague insult.
Privacy and Data Vocabulary
- personal data: information relating to identifiable people;
- consent: meaningful permission under the relevant conditions;
- data minimisation: collecting only what is necessary for the stated purpose;
- retention: how long information is kept;
- access control: who is permitted to see or use information.
A Student Verification Workflow
- Identify which parts of the output are factual claims.
- Separate explanation from evidence.
- Check important factual claims against reliable independent sources.
- Test calculations or quotations independently where relevant.
- Ask whether the answer omitted a significant counterexample or condition.
- Rewrite the final understanding in your own words.
Checkpoint: What Would Happen If the Output Were Wrong?
The higher the consequence of error, the stronger the verification and oversight should be. This connects technology vocabulary to responsibility rather than hype.
Connect this to Secondary 2 Vocabulary for Media, Influence and Information for a broader verification framework.