brightlab.Made for curious mindsBrightlab
Year 12 · Governance
Can a popular pilot still be unacceptable to the people it affects?
Go to the investigation ↓A community pilot decision record with responsibilities, dissent, appeal and pause policy.
Take this investigation into Python: open the interactive Marimo notebook →
Who can challenge this pilot and change its scope?
Governance defines who can make decisions, contest them and require change. This fictional library pilot has stakeholders with different access needs, review requirements and possible harms. Allocating review time reveals whose concerns have not been examined. The simulator reports unmet procedural constraints, not an ethical score. Majority support does not remove a minority’s privacy boundary or the need for an appeal and accountable owner.
Before this lesson: Separate technical performance, consent, representation and accountability. Useful earlier investigations: The system that fits the school.
Governance defines who can make decisions, contest them and require change. This fictional library pilot has stakeholders with different access needs, review requirements and possible harms. Allocating review time reveals whose concerns have not been examined. The simulator reports unmet procedural constraints, not an ethical score. Majority support does not remove a minority’s privacy boundary or the need for an appeal and accountable owner.
Separate technical performance, consent, representation and accountability. Useful earlier investigations: The system that fits the school.
Plan three 50-minute sessions. Assign fictional stakeholder briefs rather than asking students to represent real identities. Read access, privacy and staffing constraints. Explain that real community authority cannot be simulated or replaced by classroom votes.
ACARA V9 AI curriculum connection ↗ · Technologies ↗
A complete teaching sequence · 150 minutes
Read the library purpose and identify people affected, including those who may not use the service.
Ask: “Who can bear a cost without being a user?”
Listen for: “Staff, carers or people whose information appears in sources.”
Predict which requirements remain unmet when review time follows the largest group.
Ask: “Does the largest vote answer every rights question?”
Listen for: “No; some boundaries are not settled by popularity.”
Allocate a fixed review-time budget and inspect whose minimum review remains unmet. Set appeal and stop-rule fields and examine the process trace.
Ask: “What does an unreviewed group mean in our evidence?”
Listen for: “We do not yet know enough about its concerns.”
Test the edge case with strong overall support but an unresolved privacy objection.
Ask: “Can the model turn that objection into a small penalty and proceed?”
Listen for: “It should show the unresolved boundary for accountable judgement.”
In the third session, revise scope, review allocation and staged rollout. Assign role owners, a response deadline, accessible appeal route and a pause condition. Document unresolved disagreements in their own terms.
Ask: “What can we change about the pilot so the conflict is reduced?”
Listen for: “Scope, data collection, access options or whether we launch at all.”
Present a proceed, revise or stop recommendation with evidence and dissent recorded. Peers challenge missing representation.
Ask: “What authority would a real community need to exercise?”
Listen for: “Real participation and decision rights, not our classroom proxy.”
Majority preference removes minority harms.
A pilot can have high aggregate support while leaving a smaller group’s privacy boundary unresolved.
Produce a staged governance policy with representative review, accountable roles, an accessible appeal and explicit stop conditions.
Evaluate the reasoning and procedural completeness, not agreement with one launch decision. Require dissent and unresolved constraints to remain visible.
| Criterion | Beginning | Secure | Extending |
|---|---|---|---|
| Stakeholder reasoning | Uses only aggregate support | Identifies an underrepresented concern | Distinguishes preference, rights and decision authority |
| Governance design | Names vague oversight | Defines owner, appeal and stop rule | Tests accessibility, deadlines and unresolved dissent |
Provide role briefs and a structured decision record. Allow oral panel evidence alongside written documentation.
Design an independent review mechanism and test a case where the system owner has a conflict of interest.
The community is fictional. Do not collect real sensitive opinions or claim to speak for Aboriginal and Torres Strait Islander communities. Real cultural data require appropriate authority and co-design; local senior syllabus mapping is essential.
Shared device? Turn remembering off. A project file lets you continue on another device.
Your browser is the laboratory
Allocate review time by group size. Predict which smaller-group requirement will remain unmet before checking constraints.
A 120-minute review budget serves general visitors (need 30), access/privacy panel (need 40) and staff (need 25). Initial allocation is 70,20,30. Fictional majority support cannot override an unresolved privacy boundary.
The experiment opens after your prediction.
There is no penalty for being surprised.
Apple Silicon · PyTorch MPS
Enumerate review-time allocations using tensor operations to expose unmet procedural constraints. Change the total time budget while keeping three group requirements fixed; ethical judgement remains with people.
Students extend the experiment in teams, documenting parameters, outputs and limitations.
The bundle contains lesson-specific working code, a configuration file, a reactive notebook, a deterministic CPU check and hardware setup instructions. Acceleration is reported only after a tensor operation and result read-back succeed.
Download Mac Studio investigation ↓unzip y12-govern-mac-pathway.zip -d y12-govern-mac cd y12-govern-mac bash setup-mac.sh source .venv/bin/activate python experiment.py --device mps --output results marimo edit notebook.py
CPU and available-device execution status is recorded in the downloaded README and validation report. DGX Spark execution requires that hardware; static validation alone does not establish GPU compatibility or performance. The browser lesson remains fully available without this extension.
NVIDIA DGX Spark · PyTorch CUDA
Enumerate review-time allocations using tensor operations to expose unmet procedural constraints. Change the total time budget while keeping three group requirements fixed; ethical judgement remains with people.
Students extend the experiment in teams, documenting parameters, outputs and limitations.
The bundle contains lesson-specific working code, a configuration file, a reactive notebook, a deterministic CPU check and hardware setup instructions. Acceleration is reported only after a tensor operation and result read-back succeed.
Download DGX Spark investigation ↓unzip y12-govern-dgx-pathway.zip -d y12-govern-dgx cd y12-govern-dgx bash run-dgx.sh # Open the localhost notebook URL printed by the container.
CPU and available-device execution status is recorded in the downloaded README and validation report. DGX Spark execution requires that hardware; static validation alone does not establish GPU compatibility or performance. The browser lesson remains fully available without this extension.
A community pilot decision record with responsibilities, dissent, appeal and pause policy.
Download the editable handout →