Australia is running out of the human data that powers AI, and companies are already building systems to monetise the behavioural signals that remain. The risk is simple: institutions capture the value while consumers and the country miss out, writes Cam Partridge.
Large language models are approaching a limit.
The open web – once the fuel of generative AI – is running dry. Synthetic data can’t replicate human nuance, and reinforced learning from human feedback doesn’t scale. The missing ingredient is behavioural signal – the lived, contextual record of how people act, decide and adapt.
Enterprises are already building the rails to monetise it. The question for Australia is whether it captures any of that value – or remains a supplier of raw material while others profit.
The Data Bottleneck
The first wave of generative AI fed on the open internet – forums, code repositories, academic papers, social feeds. That source is drying up: high-quality data is finite, copyright suits are multiplying, and the performance gains from scraping more text are flattening.
To keep improving, models now need richer, high-fidelity human input – and structured ways to access it.
Enterprises Move First
The first movers are not individuals but institutions.
Companies like CarbonArc in the US are building the infrastructure to make institutional behavioural data liquid. Their ontology frameworks let disconnected datasets – from retail transactions to supply chains – interoperate through shared standards.
“Data is the largest trapped intangible asset in the world – enterprise or personal,” says CarbonArc founder Kirk McKeown. “We’re building the infrastructure to unlock it – a modular framework where datasets connect seamlessly to deliver decision-grade signals.”
This isn’t about selling personal records (McKeown notes CarbonArc operates under enterprise suppliers’ existing terms of use and upstream consents), instead, it’s about licensing de-identified, hashed insights – structured data with provenance, priced by domain, geography, and freshness.
When a retailer’s basket data can inform logistics, or a telco’s location trends can predict energy demand, signal becomes strategy. Enterprises are laying the groundwork for the next data economy – one where human behaviour, not content, becomes the scarce input.
Consumers: The Missing Market
The idea that individuals could monetise their own data isn’t new. Early efforts like Killi promised cash-for-data but never gained traction. Fetch has built a rewards ecosystem where users upload receipts to earn points from brands – effectively turning everyday purchases into monetisable data streams. None, however, have scaled meaningfully.
The barrier hasn’t been technology so much as economics: no interoperable standards, no liquidity, and few institutional buyers. That could change as AI’s appetite for human nuance grows. Once enterprises normalise the licensing of structured behavioural data, consumer participation is likely to follow.
When Data Becomes Capital
When structured and tradable, behavioural signal stops being data and becomes digital capital – richer than cookies, cleaner than trackers, more precise than surveys.
McKeown sees a world where this shifts economic models: “There’s a world in ten years where people get paid for their consumption in place of universal basic income.”
It’s provocative – but not without precedent. Artists and authors are now compensated for the training data that powers generative AI. If creative work deserves licensing fees, why not transactional decisions that shape recommendation engines, credit models and supply-chain predictions?
In that world, consent shifts from an abstract legal notion to a transaction embedded in architecture.
Australia’s Concentration Problem
Nowhere is behavioural data more concentrated than in Australia.
The big four banks handle most financial transactions. Retail giants Woolworths and Coles run loyalty programs with millions of members. Telstra, Optus and TPG hold vast communications and location datasets.
That concentration gives Australia both an advantage and a warning.
If the next data economy rewards those who can structure and license behavioural signal, incumbents are best placed to profit. But for consumers, the story is different.
The Consumer Data Right (CDR) – designed to empower them – has failed to deliver. Four years after launch, only 0.31 per cent of banking customers have used the CDR. Nearly $1.5 billion has been spent building the system, yet adoption remains negligible.
The flaw is design, not ambition. The CDR was built as competition policy, not compensation policy. Policymakers feared direct payments would lead people to “sell privacy cheaply,” so the benefits remained indirect – smoother switching, marginally cheaper deals.
Other markets are testing more tangible incentives. UK fintechs have layered cashback rewards onto open-banking APIs. South Korea is piloting marketplaces for monetising card data.
If Australia’s frameworks remain static, enterprises will license insights outward while the value – and capability – flow offshore. Yet the same concentration that limits competition could be an advantage. With the right market design, Australia could treat behavioural signal as a domestic asset class – priced, portable, and shared under models that reward participation.
Owning the Next Data Economy
In the near term, enterprises will monetise derived insights within their own ecosystems, while consumer-level wallets and permissions gradually lower friction.
Over time, signal liquidity – the ability to price, trace and trade verified human input – will define competitiveness in the digital economy.
The next great resource race won’t be for data volume, but for data with provenance and consent. Whether Australia builds that market or becomes a supplier to someone else’s will determine who owns the next decade of the digital economy.
For now, the economics are controlled by institutions. But as AI systems learn to price human signal, every consumer becomes a potential supplier. The shift from free data extraction to paid data participation could be one of the most profound redistributions of value in the digital era – if markets and regulators are brave enough to let it happen.
Cameron Partridge specialises in enterprise-scale AI capability building and GTM transformation. He has led marketing and commercial functions across global AI companies, Fortune 500 innovators, and PE- and VC-backed organisations.
Look back on the week that was with hand-picked articles from Australia and around the world. Sign up to the Forbes Australia newsletter here or become a member here.
