Fair Go Justice — Accountability in practice

OUR TRANSPARENCY
COMMITMENT

We hold the system to account. That means holding ourselves to the same standard. Here is exactly how this platform works, what AI does and doesn't do here, and how every statistic is sourced.

Why This Page Exists

GOVERNMENT-GRADE INTEGRITY

Fair Go Justice is not a backyard community website. It is positioning itself as a world-leading civil society reform platform — engaging Senators, government ministers, courts, and the public. That requires a transparency standard that matches the ambition.

This page documents every methodology decision so that any journalist, senator, researcher, or member of the public can understand exactly what they are reading and why they can trust it — or challenge it.

AI-Assisted Research — What That Means

HOW WE USE ARTIFICIAL INTELLIGENCE

Fair Go Justice uses AI-assisted research methodology as a research and synthesis tool — not as a source of truth. Every AI-assisted output is subject to the following eight-rule court-safe filter before publication:

01
Source Verification
Every statistic must trace to a government report, peer-reviewed research, or primary court record. AI does not create statistics — it finds and synthesises them.
02
No Real Names
No real individuals are named in any AI-generated content. All accounts use role-based descriptions only (e.g. "a self-represented litigant", "a public sector worker").
03
Documented Patterns
Illustrative accounts on this site are documented systemic patterns — not verbatim testimonies. They are clearly labelled as such on every page they appear.
04
No Findings of Fact
This site does not make findings of fact about any specific case. It documents systemic patterns and publicly available data.
05
No Legal Advice
Nothing on this site constitutes legal advice. Content is educational, advocacy-based, and documentary in nature.
06
Systemic Framing
All analysis refers to systemic failures — not individual wrongdoing — unless supported by verifiable public record (e.g. royal commission findings, court judgments).
07
Citation on Request
Full source citations for any statistic or data point are available on request at fairgojustice@gmail.com. We publish what we can verify.
08
No Raw AI Output
No AI output is published without human review, editing, and verification against our methodology. AI assists — humans decide what is published.
How We Actually Work

THE AI-HUMAN COLLABORATION MODEL

Fair Go Justice is built on a specific model of human-AI collaboration — not AI autonomy. Here is exactly how it works, step by step, and what each party is responsible for.

HUMAN ROLE — Bret Harding
  • ✓ Sets the research intent and policy direction
  • ✓ Approves or vetoes every piece of published content
  • ✓ Provides lived experience context that AI cannot generate
  • ✓ Verifies factual claims against primary source documents
  • ✓ Makes all editorial and strategic decisions
  • ✓ Takes legal and public accountability for all published content
AI ROLE — Research Assistant
  • → Synthesises publicly available source material
  • → Drafts content for human review and editing
  • → Structures policy arguments and economic frameworks
  • → Applies eight-rule court-safe filter to all output
  • → Produces NO findings of fact, NO real names, NO legal advice
  • → Never publishes anything without explicit human approval
THE 5-STEP WORKFLOW
1. Human sets
research brief
→
2. AI synthesises
source material
→
3. AI applies
8-rule filter
→
4. Human reviews
and approves
→
5. Human
publishes

This model is consistent with the OECD AI Principles (2024) and the Australian Government's AI Policy v2.0 mandate for human oversight of AI-assisted decision-making. It is the same model we advocate for in our AI Accountability in Justice Act 2026.

Community Stories — Methodology

DOCUMENTED PATTERNS vs REAL STORIES

The accounts on our Stories page are labelled clearly as "Documented Patterns / Illustrative Accounts". They are:

We are actively building a Verified Community Stories section. Real personal accounts submitted via our story form will be reviewed, consent-confirmed, and published separately with full transparency about their verified status. If you have a real story to share, submit it here →

Movement Supporters — Methodology

PETITION & SUPPORTER NUMBERS

The "Movement Supporters" figure reflects the number of individuals who have engaged with this campaign through petition registration, email sign-ups, social media engagement, and direct outreach responses.

We are building toward a verified digital signature system with plans to integrate with government-recognised identity verification. Until that system is operational, movement supporters should be understood as campaign reach — not a formally verified petition count.

We do not inflate, fabricate, or reset movement supporter numbers. Numbers will only increase as real engagement grows.

Statistics — Source List

WHERE OUR DATA COMES FROM

Every statistic published on this site is sourced from one of the following:

Key sourced statistics include: SRL rate tripled 4.4%→12.8% (Federal Circuit Court statistical report 2024); 73+ AI fake citations (UNSW Law, April 2026); Robodebt $548.5M settlement (Royal Commission into Robodebt, 2023); NLAP expired 30 June 2025, 388,000 CLC turn-aways (National Association of Community Legal Centres); OECD judicial integrity score 26% vs 66% average (OECD Justice at a Glance 2024).

Questions & Challenges

CHALLENGE US. WE WELCOME IT.

If you believe any content on this site is inaccurate, misattributed, or insufficiently sourced, we want to know. This platform exists to advance the truth — not to protect itself from scrutiny.

Contact us at fairgojustice@gmail.com with your specific concern, and we will respond within 5 business days with a source citation or a correction.

JOIN THE MOVEMENT →