Dieser Artikel ist derzeit auf Englisch verfügbar.
AI in the Australian Workplace: Fair Work Commission, Enterprise Agreements, and the Right to Disconnect
The Fair Work Commission is actively developing its approach to AI in workplaces. The right to disconnect, AI monitoring obligations, enterprise agreement requirements, and the FWC's emerging AI jurisprudence — the practical guide for Australian employers.
Key Takeaways
The right to disconnect (effective August 2024 for non-small business employers) creates specific implications for AI-powered workplace monitoring — monitoring systems that track after-hours connectivity may conflict with disconnect rights.
The Fair Work Commission has jurisdiction to hear disputes about AI-assisted performance management, AI-generated task allocation, and algorithmically determined remuneration — these are not technology questions, they are employment law questions.
Enterprise agreements increasingly include AI-specific provisions negotiated by unions — employers in heavily unionised sectors should expect AI governance to be a bargaining topic in upcoming enterprise agreement negotiations.
The FWC's unfair dismissal jurisdiction applies where AI-assisted performance management contributes to a termination — the Commission will assess whether the AI process was procedurally fair and whether the outcome was proportionate.
Gig economy AI governance is an active FWC area — following the 'employee-like worker' provisions of the 2023 Fair Work Act amendments, platform companies using algorithmic management face expanded FWC jurisdiction.
"Nur zu Informationszwecken. Dieser Artikel stellt keine rechtliche, regulatorische, finanzielle oder professionelle Beratung dar. Konsultieren Sie einen qualifizierten Spezialisten für spezifische Beratung."
The Fair Work Act 2009: the primary employment law framework
The Fair Work Act 2009 is the foundation of employment law for most Australian employers and employees. It governs minimum employment conditions, unfair dismissal, adverse action, enterprise bargaining, and the general protections available to employees. AI in the workplace intersects with the Fair Work Act in several specific and important ways.
General protections: adverse action and AI
Part 3-1 of the Fair Work Act prohibits adverse action against an employee because they have exercised, or propose to exercise, a workplace right. A workplace right includes the right to make a complaint or inquiry, to participate in enterprise bargaining, to take a period of leave, and to exercise rights under the Act or an industrial instrument.
AI monitoring that is used to identify employees who have complained about working conditions, or who are active in bargaining, and that leads to adverse treatment of those employees creates a general protections claim. The connection between the monitoring, the identification, and the adverse action needs to be established, but the general protections framework is a powerful tool. The employer bears the burden of proving adverse action was not taken for a prohibited reason once an employee establishes a prima facie case.
Enterprise agreement obligations and AI governance
Enterprise agreements increasingly contain provisions about technology change and, in some cases, specifically about AI. Under the Fair Work Act, employers must genuinely consult with employees and their representatives about major workplace changes — including the introduction of AI systems that significantly affect employees. This consultation obligation is not satisfied by simply announcing the change after the decision is made. Genuine consultation requires providing information about the proposed change, genuinely considering employee input, and allowing a reasonable opportunity to respond before implementation.
In enterprise agreement negotiations, unions in Australia — particularly in transport, logistics, retail, and manufacturing — have been seeking specific AI governance provisions. Model clauses being negotiated include: disclosure requirements before AI tools that affect working conditions are introduced; joint union-management AI committees with advisory rights; human review requirements for AI-generated performance assessments; restrictions on AI-based disciplinary triggers without human investigation; and transparency about algorithmic management systems and their parameters.
The right to disconnect: Fair Work Act s65A
Effective 26 August 2024 (non-small businesses) and 26 August 2025 (small businesses), Fair Work Act section 65A gives employees the right to refuse unreasonable contact outside working hours. This directly applies to AI-driven contact: automated task assignments at night, performance alerts generated outside hours, and AI scheduling tools that push urgent work after hours. The right does not prohibit employer contact — it gives employees the right to disengage from unreasonable contact without consequence.
Whether contact is "unreasonable" depends on: the reason for the contact; how disruptive it is to the employee; whether the employee is compensated for after-hours availability; the nature of the role; and the employee's personal circumstances including caring responsibilities. Disputes about the right to disconnect go to the Fair Work Commission for resolution — initially conciliation, then a binding order if unresolved.
Unfair dismissal and AI-driven performance management
The unfair dismissal framework (Part 3-2, Fair Work Act) protects eligible employees from dismissal that is harsh, unjust, or unreasonable. An employee dismissed following an AI-generated performance assessment or AI-identified misconduct has the right to apply to the Fair Work Commission if: they have served the minimum employment period (one year for small businesses, six months for others); they earn below the high-income threshold; or they are covered by a modern award or enterprise agreement.
The FWC's analysis in AI-influenced dismissal cases will examine: whether there was a valid reason for dismissal (the AI output must be reliable evidence of the reason alleged); whether the employee was notified of the reason and given an opportunity to respond; whether the employer followed a fair process; and whether the decision was proportionate. AI-generated evidence — productivity scores, monitoring data, automated misconduct detection — is not inherently reliable. The FWC can (and does) examine how AI tools work and whether their outputs were properly verified.
Modern awards and AI in schedule management
Modern awards set minimum conditions including rates of pay, penalty rates for shift and weekend work, and notice requirements for roster changes. AI-driven scheduling systems must still comply with modern award requirements. An AI that schedules workers outside the award's rostering notice requirements, or that generates schedules that avoid penalty rates in ways that conflict with award provisions, creates award compliance liability.
The Fair Work Commission reviews modern awards regularly. The 4-yearly review process and targeted variation applications provide mechanisms for awards to be updated to address AI-specific issues in scheduling, performance management, and workforce allocation. Employers and unions should monitor these proceedings in their relevant industries.
Practical compliance for employers
Ensure AI scheduling systems are configured to comply with relevant modern award rostering and notice requirements. Implement genuine consultation processes before deploying AI systems that significantly change working conditions. Review enterprise agreement change consultation clauses against your AI deployment timeline. Configure AI monitoring and performance systems to allow human review and challenge before AI outputs are used in disciplinary processes. Train HR and people managers on unfair dismissal procedural requirements — the process matters even when the substance of the dismissal reason is sound. Document AI-assisted employment decisions and the human review process applied to them.