AI Challenge

Theme: “How might we…”

Challenge Day: Thursday 15th October 2026

Overview

“How Might We…” prompts for the Trusted AI decision-support challenge pick one and run with it. Each is framed as an open “How Might We” question so you can shape your own solution, but comes with a real user, a decision to support, and the trust twist that makes it worth doing. Whatever you choose, scope it to one decision, one user, synthetic or public data — and remember the AI assists, a human decides

Tasks

1. Housing maintenance triage HMW: How might we help a housing maintenance coordinator prioritise urgent repairs across remote NT communities without “efficiency” quietly pushing remote tenants to the back of the queue?

2. Grant eligibility assessment HMW: How might we help a grants officer assess applications consistently and produce reasons a rejected applicant could actually understand and challenge?

3. Information‑access (FOI) request triage HMW: How might we help an information officer find the records relevant to a request while guaranteeing a human reviews anything sensitive before it’s released?

4. Risk‑based inspection prioritisation HMW: How might we help a regulator decide where to send limited inspectors first, without the model just re‑checking wherever it looked last time?

5. Getting urgent messages to a human, fast HMW: How might we make sure a vulnerable person’s urgent message never gets buried in a high‑volume government inbox and reaches a human quickly?

6. Making sense of long documents without over‑trusting the AI HMW: How might we help a caseworker get through a backlog of long policy or case documents faster, without nudging them into trusting a summary they haven’t verified?

Eligibility

Teams of 2–4 enrolled CDU IT coursework students (undergraduate, postgraduate, TAFE & short courses). Higher Degree by Research students are not eligible.

Evaluation

Judging Criteria

Submissions will be judged according to the following criteria: datasets, overall creativity and originality, technical sophistication, contextual relevance and practicality, ethical considerations and presentation.

First winner and runner up will be selected.

Panel of Judges

Cat Kutay

Charles Darwin University

Arafat Tehsin

Fedorai

Rushi Vyas

Open AI

TBD

To be confirmed

1. Team Up

Teams of 2–4 enrolled CDU IT coursework students (undergraduate, postgraduate, TAFE & short courses). Higher Degree by Research students are not eligible.

2. Submission

Submission items: 1. Project report (Requirement Link) 2. Source code (Python with remarks for the key steps) with a README file containing reproducing instructions. Submission Deadline: 8th October 2026 (Submission Link)