If Australia wants artificial intelligence (AI) to deliver genuine productivity gains, it should ask a question often missing from the technology debate: who is best placed to judge whether these tools make people’s lives better?
Aboriginal and Torres Strait Islander community-controlled organisations should be among the leaders of the AI revolution. Their experience offers a powerful lesson in productive, accountable, people-centred innovation.
These organisations work where administrative inefficiency has deeply human consequences. Every hour spent duplicating reports, chasing funding or navigating fragmented government systems is an hour taken from supporting families, strengthening culture and creating opportunities. AI could help, but communities must determine what better productivity means.
The pearl: more time for what matters
Australia’s National AI Plan recognises AI’s potential to lift productivity. Yet public debate too often centres on software adoption, investment and technological sophistication. The more important question is whether organisations can deliver better outcomes with available resources.
For a community-controlled health organisation, this could mean less administration and more time with patients. For a housing provider, it could mean identifying maintenance needs earlier and resolving repairs faster. For services supporting children, families or people with disability, AI could help coordinate appointments and prepare routine reports. These applications may lack glamour, but their effects could be substantial.
Community-controlled organisations navigate multiple funding agreements and reporting systems. Carefully governed AI could prepare routine documentation, flag inconsistent records, summarise non-sensitive information and make organisational knowledge easier to access. Staff would retain responsibility for decisions, professional judgement and relationships.
The aim is to remove administrative friction while strengthening the human connections at the heart of these services. Community leadership could also help Australia develop AI that is culturally safe, locally relevant and accountable to those affected.
The irritation: communities must share the benefits
Community-controlled organisations are asked to deliver increasingly complex outcomes with limited funding, overstretched staff and fragmented technology systems.
Large corporations can invest in consultants, infrastructure and experimentation. Community organisations may struggle to fund the staff and systems needed to make AI useful. Expecting adoption without supporting implementation, training, security and evaluation risks widening existing inequalities.
Imposed technology creates further risks. A system suited to a metropolitan office may be inappropriate for a remote community or culturally sensitive service. Tools trained on incomplete or unrepresentative data can produce misleading recommendations, reinforce bias and create barriers to services.
Ownership matters too. Community information is not simply a resource for technology companies to extract and commercialise. Indigenous data sovereignty requires Aboriginal and Torres Strait Islander peoples to exercise meaningful authority over how their data is collected, accessed, interpreted and used. Cultural knowledge and language demand particularly careful stewardship.
AI should strengthen community control, never quietly transfer it to a software provider, government agency or algorithm.
Every AI initiative in the community-controlled sector should meet three tests.
First, productivity created. Does it improve service delivery? Establish a baseline and track administrative hours, processing times, errors, waiting times, staff workload and service quality. Time saved must have practical value: more appointments, faster repairs, better support or less burnout.
Second, value retained domestically. Does investment build lasting capability in Australia and the community-controlled sector? Track spending with Indigenous businesses, employment, skills, locally owned intellectual property and reliance on external consultants. Organisations should have opportunities to co-design, own or govern reusable tools, rather than repeatedly purchasing solutions that leave little capability behind.
Third, benefits distributed. Who gains? Assess whether workloads become manageable, services more accessible and community outcomes better. Smaller and remote organisations must benefit alongside larger providers. Productivity gains should not simply become government savings while services remain underfunded.
Participating organisations and communities should choose these measures. Evaluation must not become another universal reporting burden.
The Australian government should fund a three-year Community-Controlled AI Productivity Partnership involving community-controlled peak bodies, participating organisations, Indigenous-led research institutions and suitable technology partners.
In year one, organisations would identify priorities, establish baselines and select low-risk applications. Funding would cover digital readiness, training, secure systems and Indigenous data governance.
In year two, organisations would expand successful applications, with Aboriginal and Torres Strait Islander staff leading implementation and sharing lessons. Participants could develop common tools where useful, with sensitive information remaining under agreed community governance.
In year three, independent evaluators would work with participating organisations to assess productivity, financial impacts, workforce outcomes and community benefits. Findings, including failures, would be published in forms approved by participating organisations. Successful approaches could then expand, backed by long-term funding for maintenance, capability and governance.
This would support the National Agreement on Closing the Gap: Priority Reform Two seeks to strengthen the community-controlled sector, while Priority Reform Four supports access to and use of data for community priorities.
Let communities define success
AI will not resolve inadequate funding, workforce shortages, fragmented government systems or the denial of self-determination. Nor should it become an excuse to cut essential services.
It could, however, help staff do their jobs better, strengthen local capability and return time and resources to communities.
Australia should invest in organisations that know their communities, give them authority and resources to experiment safely, and judge results by evidence.
The prize is a stronger community-controlled sector: less time lost to bureaucracy, greater control over information, and more resources for care, culture and opportunity.
Australia’s AI productivity revolution will be more credible if community-controlled organisations help lead it, and more just if success is measured in time returned, community control strengthened and lives improved.
The views expressed in this article may or may not reflect those of Pearls and Irritations.
