Australia’s National Disability Insurance Scheme (NDIS) will undergo major reforms in mid-2026, with participants’ funding packages and support plans to be generated by computer programs, leaving staff with no discretion to alter them.

The new model, known as the I-CAN Planning Tool (Instrument for Classification and Assessment of Support Needs, Version 6.0), was developed by the University of Melbourne and the Centre for Disability Studies, and has been used in Australia’s disability sector for the past 20 years. The National Disability Insurance Agency (NDIA) says the tool will improve plan consistency, reduce human error, and cut the cost and time participants spend gathering medical evidence.

Under the new system, the Administrative Review Tribunal (ART) will no longer be able to directly amend plans; instead, it can only send a plan back to the NDIA for reassessment. Assessors will be Level 6 employees in the Australian Public Service. Initially, assessors will be hired internally by the NDIA, and although backgrounds in allied health or lived disability experience will be considered an advantage, they are not required.

The new assessment process includes a semi-structured conversational interview and a questionnaire. When needed, targeted specialist assessments will be conducted—for example, for home modifications, assistive technology, or hospitalisation/compensation-related factors. The NDIA says delegates will be responsible for confirming whether the computer-generated plan meets a participant’s needs.

The reform will significantly reduce human involvement in plan creation and will change the appeals process. The NDIA emphasises that participants can still request a reassessment, and if they remain dissatisfied, they may appeal to the ART. However, third parties will not be allowed to intervene in the assessment process.

The Victorian government has announced it will introduce a standalone criminal offence for “coercive control” in 2026, marking a major policy reversal. Previously, the Labor government led by Premier Jacinta Allan maintained that existing family violence laws already covered coercive control and rejected opposition proposals to legislate specifically on the issue.

Last week, newly appointed opposition leader Jess Wilson made creating a coercive control offence her first commitment for the 2026 state election, promising that if the Coalition wins, the law would be introduced within her first 100 days in office. After the Liberals submitted bills to both houses of the state parliament this week, the government shifted its stance. Attorney-General Sonya Kilkenny acknowledged that “more must be done,” admitting that current laws remain insufficient to protect victims.

The government emphasized it will roll out its own legislative version in 2026, but the process will involve thorough consultation to avoid unintended legal consequences. Although Labor voted against the Liberal bill in the lower house, it supported a separate Liberal motion in the upper house to “expedite” coercive control laws.

Coercive control involves behaviors such as isolation, monitoring, gaslighting, and economic control, typically perpetrated by male offenders against female partners.

The government is also advancing another family violence reform, including setting a two-year minimum term for intervention orders, ensuring protection continues for children after they turn 18, and expanding the definition of family violence to include stalking, systemic abuse, and animal cruelty.

The Australian Labor Government announced it will not introduce a standalone AI law but will implement a National AI Plan, managing AI under existing legal frameworks.

The plan aims to balance economic growth and social equity, avoiding additional legislative burdens while accelerating domestic AI development. The government highlighted AI’s potential to improve healthcare, education, long-term care, and employment, creating jobs and enhancing technology rather than replacing labor. Public and private high-value datasets will be released for AI model training, with incentives for data center investment to boost local R&D and competitiveness.

To address risks, an AI Safety Research Institute will be established to monitor AI’s impact on energy consumption, gender misuse, and intellectual property issues, with an initial AUD 30 million investment. The government will evaluate whether deeper legislation is needed and ensure regulatory frameworks keep pace with AI development.

Some scholars and privacy advocates worry that relying on existing laws alone may lead to unclear accountability, weak oversight, and insufficient transparency. Niusha Shafiabady, an associate professor at Australian Catholic University, criticized the plan for gaps in accountability, sovereignty, security, and democratic oversight. Opponents also highlighted environmental concerns from expanding AI systems and data centers and the lack of legal protections against AI-induced discrimination, privacy violations, or manipulation.