CODE / DECISION SIMULATION / LEGAL EDUCATION
Legal Operations Portal
A live decision-making simulator for graduate legal education — teams advise under incomplete, path-dependent information while the instructor clocks the room and AI assists review under human control.
Next.js / Postgres / human-in-the-loop AI
Classification
This is simulation-based learning in a professional register — a live decision environment, not a game and not gamification. There are no points, levels, or win states. Students practise the work lawyers actually do under time, incomplete evidence, and institutional pressure: issue-spotting, authority, application, counterargument, and operational advice.
In current legal-education terms it sits with scenario-based / experiential methods and authentic assessment: performance is judged by legal craft (accuracy, interpretation, coherence, usable advice), not by proximity to a presumed doctrinal mainstream. AI is used as instructor-side decision support under entitlement rules — human-in-the-loop, never the author of the factual world.
What runs in the room
Student teams advise a fictional state through timed live sessions (map and overlays, live feed, evidence, world materials, workspace, COMMS, compulsory legal tasks). Information is path-dependent: teams share a common world but not the same entitlements. The instructor clocks the block — live → freeze → processing → review → carry-forward — so pedagogy and fairness stay under human control between autonomous hours.
Built for Use of Force (Melbourne Law Masters, LAWS70078, October 2026). Session 1 is the first playable slice; further sessions are in development.
What is assessed
The simulator scores whether a team can construct and defend a legally serious position from the materials available to it — identifying governing issues, using authorities according to their status, applying material facts to the stated test, confronting counterarguments, and giving advice that can be acted on. A restrictive, expansive, or reserved position can score equally well if the craft holds. Unlocking more favourable path information is not itself a mark of quality.
AI role
Bounded instructor-side roles (query interpretation, information composition within entitlements, reaction and position extraction, rubric assistance). AI is never WORLD_AUTHOR. Student-visible factual outputs must satisfy entitlement and validation checks; the instructor retains start, pause, freeze, and carry-forward.
Public status
No public login or repository is linked here yet. Accounts and classroom runs stay instructor-controlled for Melbourne LAWS70078. When a public demo or open materials package is cleared, links will appear on this page.