Beyond blanket safety: a risk‑led response to shifting Safe Work rules

Safe Work rules are shifting: treat compliance as an operating system

Debate over “blanket” Safe Work rules has flared, with miners and farmers warning that one‑size‑fits‑all obligations don’t fit diverse hazard profiles. That argument is a useful prompt for every executive: regulation is tightening and expectations are shifting, regardless of industry. Treat this moment as a signal to modernise safety management rather than resist it. The goal is not looser compliance, but smarter compliance—designed around real risks, dynamic work and supply chains that stretch well beyond the site fence.

A risk‑led approach starts by mapping material hazards and the contexts in which they arise, not by padding manuals. Build a plain‑English risk taxonomy, then link each category to specific controls and accountabilities. Prioritise scenarios where energy, movement, isolation or exposure combine to create catastrophic potential. This helps leaders move from generic procedures to targeted protections that travel with the work—whether it’s a mobile crew, a contractor with different systems, or a digital process that quietly changes overnight.

The business case is straightforward. When safety controls are tuned to real work, they reduce stoppages, rework and costly disputes, while demonstrating diligence to regulators and boards. Current reporting on regulatory and technological churn is a reminder that “set and forget” compliance is no longer credible. The organisations pulling ahead are institutionalising learning, not just rulebooks: they collect quality data, test controls against failure, and make it easy for frontline workers to surface weak signals before they become incidents.

Work Report note · News analysis · Current-news analysis

Build a risk‑led safety engine

Start with people. Nominate named risk owners for each critical hazard, ideally in the line, not just in safety. Equip supervisors with competency‑based training that blends hazard identification, coaching and incident decision‑making under pressure. Give Health and Safety Representatives real influence through co‑design sessions and access to data, not ceremonial roles. Close the contractor gap by onboarding them to the same risk language and escalation pathways, so accountabilities don’t dissolve at the gate or in the field.

Then fix the processes. Create a tiered control library that distinguishes critical controls from housekeeping. Bake short, scenario‑based pre‑start reviews into the rhythm of work, instead of bolting on long checklists. Establish change‑management triggers for when equipment, substances, software or shift patterns change, ensuring risk assessments update with the work. Use after‑action reviews to test whether controls actually arrested the energy or exposure in play, and capture learnings in a reusable way.

Finally, organise information. Build a single source of truth for hazards, incidents, near misses and improvement actions, with rock‑solid audit trails. Make it mobile‑first so crews can log observations and photos quickly, online or offline. Use simple dashboards that highlight drift—open actions ageing, controls not verified, areas with rising exposure—so managers can intervene early. Integrate this with existing WHS systems to avoid double handling, but don’t let legacy tools become an excuse for fuzzy, delayed data.

Work Report note · News analysis · Current-news analysis

Treat the Safe Work debate as a catalyst. Build a risk‑led safety engine—people, process and information—then add carefully governed AI and a tight 90‑day plan.

Using AI carefully under evolving Safe Work rules

AI can help, but only under disciplined governance. Sensible applications include computer vision to flag missing PPE in live feeds, or natural‑language tools that surface patterns across thousands of incident narratives. Start with contained pilots and clear success criteria; evaluate false positives alongside genuine finds. Document the workflow so humans remain accountable for decisions. Treat AI as a power tool for risk detection and analysis, not an oracle, and explain that distinction explicitly to leaders and crews.

Governance matters because expectations are changing. As government consults on the future of AI, transparency, accountability and human oversight are front of mind for policymakers and boards. Create lightweight model factsheets, record data lineage, and define who reviews, overrides and escalates algorithmic outputs. Keep safety committees close to trials and make privacy impact assessments routine. Those practices translate regulator intent into pragmatic guardrails, and they build confidence that AI won’t become an unreviewed “second system” at work.

Worker trust will make or break adoption. Involve Health and Safety Representatives early, run explain‑back sessions, and publish plain‑English FAQs on what systems do and don’t do. Follow surveillance notification requirements and avoid any perception of secret monitoring. Make it policy that no disciplinary decisions are taken on AI outputs alone, and train supervisors to interpret alerts as hypotheses to investigate. That signals respect, improves data quality and keeps the emphasis where it belongs—on controlling real risks.

Work Report note · News analysis · Current-news analysis

A 90‑day plan to move beyond blanket compliance

Days 1–30: convene cross‑functional leads and map your top five critical hazards by work context. Run a rapid maturity diagnostic across people, process and information, and measure lead indicators like verification rates, not only lag. Co‑design new control standards with supervisors and HSRs, and nominate accountable owners. Produce a short decision paper for the board outlining risks, proposed guardrails and budget envelopes. Pick two areas with clear exposure to trial targeted controls and rapid‑capture reporting.

Days 31–60: launch the trials. Stand up daily learning loops that review observations, near misses and control verifications, and tune workflows weekly. Implement minimally viable tooling—often a mobile form and a dashboard is enough at first—while scoping longer‑term integrations. Define data retention, access and de‑identification rules up front. If AI is in scope, document the pilot brief, guardrails and opt‑out choices; publish a plain‑English notice to crews and offer office‑hours for questions.

Days 61–90: scale what works and retire what does not. Tighten governance by establishing a cross‑functional control integrity council, and schedule quarterly deep dives into two critical risks. Prepare for external scrutiny by rehearsing how you evidence oversight, verification and learning. Codify your approach in a short safety charter and supplier addendum so contractors align. Close the loop with crews by sharing what changed because of their input, and the concrete hazards now better controlled.

Work Report note · News analysis · Current-news analysis