This is part two of a four-part article series. Read part one here: Universal Basic Intelligence (UBIQ): Why AI Should Be Like Your Local Public School.
When Australian political and corporate leaders talk about artificial intelligence, they inevitably speak about the “frontier.” They talk about improved productivity, world-leading models, massive data centres, and high-growth firms competing on the global stage.
This narrative leaves out most of the country. It also risks misunderstanding the nature of the challenge. The fundamental divide emerging in the age of AI is not primarily a productivity divide. It is an intelligence divide. The people, communities, and organisations that learn to collaborate effectively with machine intelligence will gain new forms of agency. Those that do not risk having their lives determined by power structures, knowledge, and systems they don’t understand and cannot participate in. While Silicon Valley chases enterprise efficiency and Brussels crafts defensive regulatory safeguards, Australia’s policy debate risks falling into the trap of designing systems primarily for the digitally privileged and white-collar workforce, while treating others as an afterthought.
If we continue down this path, AI will exacerbate our societal divides. It will give those who already possess cultural and financial capital more leverage over future forms of collective intelligence while pushing some groups – such as regional communities, First Nations peoples, elderly citizens, linguistically diverse and lower-income workers – further to the margins.
As we navigate this shift, we must guard against a narrow view of intelligence. AI risks flattening the human experience if we lose sight of the value of tacit knowledge, vocational craft, and deep cultural customs. These forms of knowledge may not be easily digitised and ingested by major AI training systems, yet they form the bedrock of communities.
To prevent AI from becoming an instrument of inequality, Australia must shift its focus beyond productivity tools and look to the tail of technology adoption. Our AI narrative needs to address a fundamentally different question: How do we build community capability that lifts the floor for everyone, rather than just raising the ceiling for a few?
Upending the natural distribution of intelligence
Historically, human intelligence and potential have been roughly evenly distributed across the population at birth. While systemic gaps in health, housing, and wealth have always skewed life outcomes, public systems—most notably universal public education and health services—have helped bright children from disadvantaged backgrounds still have a genuine shot at building agency and achieving mobility.
An AI-augmented world completely upends this starting point.
In a world integrated with AI, cognitive capability is no longer just a function of human starting point and effort. It becomes a compound dynamic: starting cognitive potential multiplied by the sophistication, tuning, and access to personalised AI and agents.
If high-performing, personalised AI co-pilots are accessible only to affluent, urban, and highly educated individuals, advantage will compound exponentially. A laissez-faire approach to AI does not just reproduce existing inequalities—it creates a challenging cognitive divide.
Just as the rationale for universal public education was built on preventing wealth from dictating a child’s life outcomes, effective pre-distribution of AI is required to ensure that technological augmentation does not become a luxury good reserved for the few.
AI access is not AI agency
The default administrative response to technological change is transactional: roll out fibre-optic networks, distribute discounted tablets, stand up an informational website or run short-term webinars on basic computer literacy. While physical connectivity remains an essential prerequisite, providing access to an AI chat interface is not the same as giving a community control over its digital future.
Dropping an AI tool into a regional community without the capability to prompt, challenge, or adapt it is like leaving a high-tech medical scanner at a regional hospital without a technician to operate it, and then declaring the health care divide addressed.
Australia doesn’t just face a digital access gap; we face an agency gap. In practical terms, this means the difference between using AI to complete a task and using AI to expand the range of possibilities available to a person or community – including non-commercial possibilities, such as engaging with family, exploring a cultural past or dreaming of a different future. One is productivity. The other is agency.
When citizens are passive consumers of offshore commercial models, they conform to assumptions baked in elsewhere. Real equity requires more than a login to a tool. It takes shape when communities have the skills, local knowledge, and infrastructure to shape, co-create, and direct these tools on their own terms. The goal is not just digital inclusion alone. The goal is meaningful participation in future intelligence systems.
Who is currently being left out?
When we look beyond the metropolitan corporate landscape, the risk of exclusion manifests across several critical domains:
- Cultural knowledge and indigenous sovereignty: Generative AI models are fundamentally pattern-matching engines trained on vast, digitised swathes of the Western (largely English-language) internet. Knowledge systems that are non-digitised, oral, place-based, or relational—such as First Nations songlines, deep environmental management practices, and Indigenous governance—are structurally excluded. Without deliberate, community-led intervention, AI risks flattening cultural diversity and crowding out alternative ways of knowing.
- Regional, rural, and agricultural communities: Regional Australians already know what it feels like to be on the receiving end of centralised, automated decision-making. When public AI is deployed through a narrow administrative lens, rural communities risk having their public services managed by distant algorithms—from automated compliance flagging on income support to formula-driven cuts in local road funding or health staffing. Without local nuance, human recourse, or community context built into these models, efficiency at the centre gets neglected at the perimeter.
- Non-white-collar workers and the care economy: While office workers build AI capability during paid desk time—and deploy those same tools to manage their personal life admin—those who work on their feet do not get that opportunity. An aged care worker or a tradesperson isn’t given an enterprise license or time to experiment at work. They may miss out on using AI for the complex, time-consuming tasks outside of work, whether that’s navigating government support systems, challenging an unfair bill, or planning their family’s finances. When AI capability is only a byproduct of desk work, we deepen the divide between those who have the tools to navigate modern life efficiently and those who must carry the manual administrative burden alone.
- Elderly citizens and migrant communities: If government and commercial services default to AI-driven chatbots without community-based support, citizens with limited English proficiency or lower digital confidence can be systematically shut out. AI prowess becomes a form of “cultural capital”—a barrier that determines who can navigate state systems and who gets stranded in automated phone queues.
Beyond the NDIS thinking: The limits of state provision
How should government respond to this challenge?
The traditional reflex is to create an administrative entitlement—a government program that assesses individual eligibility, hands out personal vouchers, and relies on a market of providers to deliver services.
We have seen the structural limits of this approach in the administrative challenges surrounding Australia’s National Disability Insurance Scheme (NDIS). If government positions itself as a funder of individual entitlements, it can inadvertently:
- Disempower social and community support networks in favour of marketised service providers.
- Create bureaucratic complexity that privileges those who know how to play the system.
- Crowd in bad actors that seek to exploit systems for personal gain rather than social benefit.
- Trap public administration in compliance auditing rather than fostering organic capability
- Leave high ongoing fiscal liabilities for current and future taxpayers.
Government cannot—and should not—attempt to be the sole provider of community AI agency. A government department in Canberra or Melbourne cannot teach a neighbourhood house in regional Queensland or a remote landcare group how to apply AI to their specific local challenges.
The role of community anchors and family networks
If top-down government webinars and market entitlements won’t build genuine agency, how does capability grow?
It grows through relational social infrastructure.
Just as children learn their most vital tacit knowledge, values, and cultural norms from parents, extended family, and community networks—long before they set foot in an early childhood centre or classroom—adults build digital capability through trusted, local relationships.
Internationally, forward-thinking initiatives demonstrate what this looks like in practice. One such initiative is the Google.org’s AI Opportunity Fund delivered by CPI, where civil society organisations and local libraries are equipped to deliver tailored, high-touch AI capability directly to underserved populations, including migrant workers, carers, and older citizens.
Rather than imposing a government-centric channel, these models fund trusted local intermediaries who already have relationships with the community.
What must government do?
To ensure no community is left behind in Australia’s AI transition, the state’s mindset must quickly move beyond being a service provider to an active facilitator and funder of community capability. We do not need a tech-funded training scheme, nor a market-based subsidy to enable access to AI for the least well-off. We need systemic shifts which move AI, power and agency into community settings.
To operationalise Universal Basic Intelligence (UBIQ), governments across Australia should commit to four structural shifts:
- Support community anchors: Direct public investment and ongoing capability building to existing local infrastructure—public libraries, neighbourhood houses, First Nations community organisations, and regional hubs—enabling them to act as physical spaces and relational networks for peer-to-peer AI learning and co-design (alongside their other valuable activities). Many are already engaged in digital inclusion activities; they need ongoing support to shift into higher-order AI co-intelligence partners.
- Prioritise the “tail” over the “frontier”: Evaluate public technology investments and national AI strategy not by how they accelerate office worker efficiency, but by how effectively they improve social participation, civic agency, and social cohesion for the most disadvantaged in the population.
- Protect sovereign knowledge commons: Establish public data trusts and sovereign technical architecture that allow local communities and First Nations custodians to digitise, hold, and control their own knowledge without surrendering IP to global technology firms, aligned to the work of organisations like Old Ways, New. We need communities to write themselves into the stories of our culture and society (and yes, our economies).
- Embrace Relational Public Administration: Replace short-term competitive grant cycles with long-term, relational models that allow community organisations to innovate, experiment, and build co-intelligence at their own pace and on their own terms.
Building a co-intelligent society for all
The choice facing Australia is not whether AI will take a central role in our society—that transition is already well underway. The choice is whether we see machine intelligence primarily as a commercial tool that can inadvertently deepen inequality, or whether we treat it as a universal public endeavour that expands human agency and brings us all closer together.
A co-intelligent society is not one where everyone uses AI. It is one where everyone has the opportunity to contribute to, shape, and benefit from new forms of collective intelligence. By focusing on community capability, respecting diverse ways of knowing, and investing in trusted local institutions, Australia can ensure that co-intelligence becomes a public good rather than a privilege.
In our next article, we will examine the intellectual heart of this journey: how human wisdom, creativity, and Indigenous knowledge systems can partner with machine capability through the practice of co-intelligence.
