Featured this week
Start here.
Six Practices Replaced Ten Guardrails. Remap Now.
The National AI Centre published six essential practices on 21 October 2025, and the ten guardrails an Australian control matrix written in late 2024 was keyed to are now superseded. The official mapping is clause level and many to many, several targets do not resolve, and the fix is a rebuilt crosswalk, not a find and replace.
The Ride Had No Way to Stop Itself
The Coroners Court of Queensland found that a basic automated water level detector and a single emergency stop were both missing, and named those absences as deficiencies. That inverts the reflex that a person at every decision point is the safe design, and it sets the test for every AI and sensing control.
Read article
Your Claim Risk Just Went Up Forty Percent
The Fair Work Commission recorded 53,617 lodgments in FY2025-26, unfair dismissal up 40 percent on FY2020-21 and general protections disputes up further again. From 20 October 2026 an applicant who used generative AI must disclose that use, and every applicant must declare the form has been checked. Here is how to re-baseline the dispute provision and the response roster.
Read article
Which of Your Prompts Are Records Now
Most prompt and response logs are facilitative and disposable. A few are the only evidence of how a business record was made, and none of the four assistants read for this article separates the two for you. Australian Privacy Principle 11.2 pushes against keeping everything, while vendor defaults run from an 18 month auto-delete to indefinite. Build the trail instead.
Read article
Gemini 4 Argon Is Out. Test It on Your Own Work First
Google's Gemini 4 Argon is generally available and pitched at legal and finance work. Its own model card shows it behind at least one rival on 5 of the 19 benchmark rows it publishes, in a table Google compiled from its own runs, public leaderboards and providers' reported scores. The result that counts is the one on your tasks, so run a three-step test before you switch.
Read article
Write the Consultation Pack Before the Tool Arrives
An AI tool that changes how a team works is a health and safety matter under section 49(d) of the model WHS Act, so the duty to consult in section 47(1) falls due before the tool lands. The artefact is a four part pack drafted from the rollout plan and role descriptions, with the leader writing the decisions.
Read article
Your Privacy Policy Now Has to Name the Decisions
From 10 December 2026 an APP entity's privacy policy has to set out the kinds of personal information its computer programs use, and the kinds of decisions those programs make or shape, where the decision could reasonably be expected to significantly affect rights or interests. In the Comcare scheme that reaches the assistive layer, as at 26 September 2026.
Read articleSections
Pick your starting point.
AI News
The latest model releases, policy developments, and industry shifts, curated for professionals.
WC & AI
AI applications in workers compensation under the SRC Act and Comcare scheme.
GRC
AI governance and regulatory developments for compliance and risk professionals.
Learning Hub
Practical guides, tutorials, and deep-dive courses on the Acts and standards that govern your work.
AI comparison table
The leading AI tools, scored honestly.
Four highlights below; the full table scores all five tools, including Microsoft Copilot, across six criteria with role-specific lenses. Updated when something actually changes, not every time someone ships a demo.
Anthropic
OpenAI
Perplexity AI
AI Readiness Assessment
Find out where you actually stand, and exactly what to do next.
Two ways in. Take the quick check for an at-a-glance read on your role's AI maturity, or unlock the full assessment for a deeper diagnostic, a 90-day plan, and a Personalised Prompt Pack.
Quick check (15 questions)
A 5 minute pulse on your role's AI readiness. Tier badge, one to two prompts, downloadable PDF. No email needed.
Start the quick check →Free, deeper diveFull assessment (~25 questions)
A complete diagnostic across four dimensions. Radar chart, 90 day action plan, three to five prompt Personalised Prompt Pack with full setup, risk and rubric notes.
Start the full assessment →Our method
Five steps between an AI draft and a decision.
TheAICommand works to the Verified Draft Method: de-identify the inputs, ground the model in your own source material, keep a person at the decision point, verify against the primary source, and log what happened.
Workers Compensation & AI
Where AI meets the SRC Act, written peer-to-peer.
Practical guides, governance explainers, and ART case-law briefings. Written peer-to-peer for case managers, claims officers, and scheme advisors. Educational only, never legal advice.
Leveraging AI to assist dissecting the SRC Act Review.
A practitioner workflow for employers to dissect the December 2025 SRC Act Review with AI tooling, end to end.
The De-Identification Toolkit for Case Managers Working With AI.
A practical de-identification framework for claim narratives before any retrieval call leaves your perimeter.
Take the WC Self-Assessment
Map your SRC Act practice capability, with AI leverage woven through. An educational self-reflection tool with a personalised PDF report: radar chart, 90-day plan, prompt pack, and risk heatmap.