Intelligent systems & ethics
When automation helps — and when humans must decide
When should an intelligent system decide — and who bears the consequences if it is wrong?
Technology inquiry
Intelligent systems & ethics
When automation helps — and when humans must decide
Digital Technologies · Intelligent systems · Ethics
Wero
When should an intelligent system decide — and who bears the consequences if it is wrong?
First step
List stakeholders and harms; test the system on three cases including one harmful edge case; log overrides needed.
What you will show
Ethics brief with human-override rules, data minimisation plan, and named accountability.
Local place context
Which AI or sensor system are you critiquing — name data sources and limitations.
Digital Technologies · Intelligent systems · Ethics
First step
List stakeholders and harms; test the system on three cases including one harmful edge case; log overrides needed.
Expected outcome
Ethics brief with human-override rules, data minimisation plan, and named accountability.
You will investigate a real AI tool used in science, compare its outputs to human expertise, and argue for fairer or more careful use. Sources of bias, error, and limitation in an AI system applied to a science or environmental question. Bias investigation notes, human vs AI comparison tables, ethics rubric scores, and a speculative design sketch.
Five ways you could investigate
Pick one to start — or write your own question. The AI mentor supports you gently inside your investigation.
Idea 1
Training data bias
What groups are over- or under-represented in your examples?
Start with this question →Idea 2
False positive cost
What harm if the system says yes when it should say no?
Start with this question →Idea 3
Human override need
When must a human review the AI output — evidence from tests?
Start with this question →Idea 4
Transparency test
Can you explain why the system gave one answer?
Start with this question →Idea 5
Data minimisation
What data can you remove and still meet the purpose?
Start with this question →
Five things you could build
Fabrication ideas linked to makerspace tools — 3D print, laser cut, Arduino, data products, and more.
Build 1
Ethics decision flowchart
Poster: when to use, pause, or stop the system.
Open in outcome selector →Build 2
Demo interface mock-up
Show human-in-the-loop review step.
Open in outcome selector →Build 3
Sensor guard mount
Physical guard that limits what the system can sense.
Open in outcome selector →Build 4
Community briefing
One-page plain-language summary for whānau.
Open in outcome selector →Build 5
Logging Arduino sketch
Log when the model fires and with what inputs.
Open in outcome selector →
AI mentor (inside your investigation)
No separate mentor page — support appears in your investigation workspace. It starts gentle: short prompts about your research context, data, and analysis. You or your teacher can turn assistance off for unassisted work, or request more help when you need it. It also guides fabrication choices tied to your evidence.