Regulation of artificial intelligence (AI) has massive implications for health and healthcare, and health leaders must step up to fill gaps in the Federal Government’s landmark announcement this week.
Melissa Sweet writes:
The health sector – especially people and organisations with expertise in health in all policies and public health – has been urged to engage actively with the Federal Government’s plans to develop national standards for artificial intelligence (AI).
Prime Minister Anthony Albanese announced this week that the standards would “set clear rules for large data centres”, including where they are built, and around their energy and water use. He plans to introduce related legislation to Parliament early next year, following negotiations with Premiers and Chief Ministers at National Cabinet.
Health leaders told Croakey the standards should include a focus on the determinants of health, health equity, healthcare and environmental justice. They noted that the Prime Minister did not address health issues in his speech this week – despite referring to Medicare as an example of Australian policy innovation.
Nor does the Office of AI, established in the Department of Prime Minister and Cabinet on 15 July, mention health issues, or engagement with Health Ministers or departments.
Senior medical leader Professor Nick Talley welcomed the PM’s speech, while noting that it was short on detail, especially on key public health concerns such as water usage and regulation of misinformation and disinformation.
Talley urged organisations such as the National Health & Medical Research Council, the Australian Academy of Health and Medical Sciences, and the Australian Academy of Science to “step up” and engage the Government in the significant health issues involved.
While AI could bring benefits, there were also significant and wide-ranging risks, such as concerns for people losing their jobs, including in the health sector. “AI is rolling out whether we like it or not, and we need to recognise that; it’s just the regulation needs to be tight,” he told Croakey.
Talley said policy silos made it difficult for health considerations to be integrated into policy development across other portfolios, and supported the idea of the Government convening a national summit to hear from across civil society, including the health sector, about how AI regulation could be framed to ensure maximium benefits and maximum protections.
For example, in a time of housing crisis, perhaps AI companies could be required to invest significantly in housing for communities where data centres are to be located, he suggested.
Associate Professor Peter Sainsbury, a retired public health worker and Pearls & Irritations columnist, noted the “glaring” absence from the PM’s speech of “any mention of how AI will affect people’s health, health services, the environment – apart from vague comments about water and power – [and] disadvantaged people in general”.
Sainsbury stressed the need for planning to address the inevitable job losses, and asked: “How is AI going to be used to improve the daily lives of the various groups of disadvantaged people in Australia, not just the super-rich and the middle classes? How will AI contribute to reducing inequality in Australia?
“If AI is going to benefit Australians, these issues must be tackled right from the start, not be put aside for later. As far as the disadvantaged and equity are concerned, ‘later’ almost always means ‘never’ in Australia,” he said.
Also noting the PM’s lack of focus on health, Dr Tarun Weeramanthri AM, Adjunct Professor, School of Population and Global Health, University of Western Australia, said “it is now up to our sector to propose how we might contribute to the framework”.
This was especially important “given the sensitive data privacy issues around health data and use of AI, and the existing regulatory responsibilities of public health agencies, at a state and territory level, for safeguarding water supplies”, he said.
“Certainly, there is an opportunity for the sector, led by government, to create health-specific standards, that could align with work being undertaken internationally – for example through the Global Initiative on AI for Health.”
Weeramanthri, who has previously urged the public health sector to engage with Australia’s social media ban for children, also suggested the creation of health-specific governance bodies to oversee the use of health information within AI platforms and tools.
“And with a promised Digital Duty of Care likely to be introduced in coming years, the health sector has a chance to outline the responsibilities technology companies have to provide products that are actually healthy – for both mind and body – in both social media and AI fields,” he said.
Sustainability researcher Rachel De Sain said it was important the public health community engages with AI regulatory development “as proactive, collaborative strategic partners”.
“The Prime Minister’s desire to ‘shape the future’ gives us the perfect opening to ensure that digital infrastructure planning, economic frameworks, and health strategies are pulling in the same direction,” she told Croakey.
“We should actively participate in policy design to ensure that environmental impact statements for AI infrastructure include downstream health metrics.”
Public health researcher Tan Nguyen told Croakey that while AI had demonstrated useful applications in many areas including healthcare, there is a broader role for the public health community in promoting AI literacy, and its safe and effective use, while considering the real trade-offs in climate health impacts.
Healthcare matters
Consumers Health Forum of Australia CEO Dr Elizabeth Deveny said it was striking that the PM’s speech began with a reference to Medicare as an example of Australian policy innovation but then “said so little about what AI will mean for health and healthcare”.
“We welcome national leadership on AI, but health is a high-risk setting. We heard a lot about the power, water and infrastructure needed for data centres, and much less about the safeguards needed when AI affects someone’s diagnosis, treatment or care,” Deveny said in a statement to Croakey.
“Health consumers need to be involved from the start, not brought in once all the important decisions have already been made. People should know when AI is being used, what happens to their health information, what evidence shows a tool is safe, and who is accountable when something goes wrong.
“There is also a very real equity issue here. Every time we digitise healthcare, we risk making things easier for people who are already well connected and harder for people who are not. That includes people with poor internet access, low digital confidence, disability, language barriers or limited access to devices.”
If AI does deliver genuine productivity gains, Deveny said health consumers should see the benefit. “The savings should help fund better access, more human support and better care for people who are most likely to be left behind,” she said.
“AI might make parts of healthcare faster and easier. Good. But faster is not the same as fairer, and efficiency on its own is not enough.”
Meanwhile, the Royal Australasian College of Physicians (RACP) today issued a position statement warning of regulatory gaps around AI in the health system, which left doctors exposed to being held responsible for patient harm caused by AI.
The position statement calls for indemnity, contractual and insurance arrangements that fairly apportion liability risk between model developers, vendors, employers, service providers (like hospitals and clinics), clinicians and users of AI. The College also recommends that clinicians should only use AI tools that have been reviewed by the TGA (or equivalent).
A discussion paper released earlier this year by the World Health Organization considers the potential for AI to support improvements in evidence informed policy making but also notes many areas of concern, including regulatory gaps, lack of transparency, potential for bias and discrimination and the market power of a few commercial actors that dominate AI development.
“AI systems risk perpetuating epistemic injustice by privileging quantifiable, data-rich evidence from dominant institutions while marginalising other valuable knowledge sources such as lived experience, local expertise, Indigenous knowledge systems and community-based insights,” said the paper.
Publications by the Lowitja Institute and CSIRO have highlighted particular concerns for Aboriginal and Torres Strait Islander people about AI in healthcare.
While it brought opportunities to strengthen prevention, diagnosis, service planning, population health monitoring, and culturally responsive models of care, across urban, regional and remote setting, the Lowitja paper said these benefits are not guaranteed and depend on how AI systems are designed, governed, and deployed.
“Without appropriate governance, ethical safeguards, workforce capability, and culturally grounded policy frameworks, AI also has the potential to entrench existing inequities, reinforce bias in clinical and non-clinical decision-making, widen the digital divide, and reproduce or amplify structural and colonial harms within the health system,” said the Lowitja Institute’s scoping and policy review.
It calls for Aboriginal and Torres Strait Islander leadership, perspectives, and Indigenous Knowledge Systems to be embedded throughout the AI lifecycle, including ensuring that development and implementation are guided by principles of self-determination, cultural safety, Indigenous Data Sovereignty, and community control.
“For Aboriginal and Torres Strait Islander peoples, communities, and organisations, preparedness for AI is no longer a future consideration, it is an immediate priority,” says the paper.
Behind the eight-ball
Some experts approached by Croakey for comment on the Government’s plans to regulate AI suggested the public health sector had been slow to engage with the fast-moving field, although the Public Health Association of Australia is organising a conference on AI for November.
“AI is clearly a vital public health issue and one in which I fear public health is behind the eight-ball in terms of looking at the issue through the lens of the commercial determinants of health,” said health equity expert Professor Fran Baum AO, from the University of Adelaide.
“The field of AI is heavily privatised and under-regulated. There are also huge areas of uncertainty about its impact on human society and so on human health. The PM’s initiative is welcome as a first step.
“It is vital that the mental and physical health impacts of AI and its required infrastructure are rapidly studied, documented and then acted upon and that the Government’s determination to regulate is realised.”
Monash University researcher Associate Professor Zerina Lokmic-Tomkins told Croakey she was concerned that much of the public health community and also social scientists had overlooked concerns surrounding AI.
“To be honest, I am worried by how relatively quiet the public health community has been in the debate around AI. Much of the discussion has been led by computer scientists, economists, lawyers, and industry, despite AI having profound implications for the social, environmental, and commercial determinants of health, as well as for health equity,” she said.
“Public health professionals should not be in a position where AI is simply thrust upon us and we are left responding only after it has been deployed. We need a stronger and more visible public health voice to help shape how AI is designed, implemented, regulated, and governed so that it advances, rather than undermines, population health and planetary health.”
Lokmic-Tomkins said that when she proposes speaking about the environmental impacts of digital health, she is often asked whether she could instead talk about topics such as single-use plastics.
“Are we avoiding the uncomfortable reality that digital health, including AI, is not a panacea for the climate or planetary health crisis and that there is a substantial environmental costs at a time when we live in the sixth mass extinction driven by human activity through unsustainable use of land, water, energy and climate change?”
Climate crisis as he spoke
As the PM was preparing and delivering his speech, headlines around the world told of the escalating toll of the climate crisis, from the dangers of bushfire smoke across north America to lethal fires in Spain, and the diversion of finance from climate initiatives to AI.
In addition to its well-known impacts upon energy, water and resources, the AI juggernaut involves an enormous opportunity cost – diverting policy attention and investment from climate action.
How different might the PM’s speech have been if it was framed by the imperative for climate mitigation and adaptation?
Lokmic-Tomkins would have liked to have heard from the PM a broader vision of what Australia hopes AI will achieve beyond economic productivity and sovereign capability.
“While the speech rightly emphasised innovation, investment and national competitiveness, I see these as means rather than ends,” she said.
“The ultimate purpose of AI should be to improve the wellbeing of our society while safeguarding the environmental systems upon which the health of all living organisms depends.
“Stronger commitments to independent evaluation of AI’s social, health and environmental impacts, investment in public-interest AI research, AI literacy across the community, and ensuring rural, regional and disadvantaged communities benefit equally from AI is needed.
“Nothing was said on transparent reporting of the environmental footprint of AI infrastructure throughout its lifecycle, embedding public and community voices into AI governance, rather than limiting consultation primarily to industry.”
Medical leader Professor Jennifer Martin said she would have liked to hear the PM talk about how AI could help with mitigating climate change, and plans to ensure AI was investing in transition communities, such as the Hunter Valley. She also wanted to hear about using AI to create alternate jobs in advanced manufacturing, for example of medicines and IV fluids.
Lokmic-Tomkins would have liked to hear the PM say that AI policy should be guided by the precautionary principle.
“We have repeatedly seen digital technologies deployed first and governed later,” she said. “AI is troubling because it is developing at such an alarming rate and we are rushing to ubiquitously i implement it without having guardrails that ensure that critical thinking and reasoning remain intact and that societal cohesion and social equity are pillars of working with AI.”
She said AI should be evaluated by its lifecycle of greenhouse gas emissions, water consumption, resource use, impacts on biodiversity and contribution to health equity.
“We need to highlight that AI is harming planetary health and not just by contributing to climate change. AI is in the end a physical system that depends on energy, water, land, critical minerals mined from Earth and digital infrastructure. Every stage of its lifecycle has environmental consequences that ultimately affect human health.
“While the Prime Minister acknowledged the energy and water demands of data centres, the broader environmental footprint of AI received little attention. As AI becomes ubiquitous, it will need growing amounts of electricity, freshwater for cooling, mineral extraction, land development and rapidly expanding digital infrastructure, all of which contribute to greenhouse gas emissions, electronic waste, biodiversity loss and pressure on finite natural resources.
“These environmental concerns are public health concerns because they influence climate change, water security, food systems and the conditions that enable communities to thrive. They should be part of AI regulation.”
We also need to ask what AI is being developed and used for, as not every AI application delivers the same public value, she said.
“From a planetary health perspective, priority should be given to AI that demonstrably improves health, climate resilience, disaster preparedness, education and social wellbeing, rather than simply maximising commercial productivity.
“What we are seeing at the moment is that every time we open a browser AI is ready to search our query – most people will have no idea that they can switch that off, so we burn planetary resources for something as simple as to identify a meaning of a word.
Rachel De Sain added: “At a time when the climate crisis demands that we carefully manage our total environmental footprint, it is not just about the physical footprint of the data centres and their natural inputs – water, energy, critical minerals – or the direct impacts to nature like noise and light pollution affecting pollinators. It is fundamentally about what the systems being built are used for, and the indirect public health harms that flow from that.”
De Sain also highlighted the importance of protecting health data: “While the PM spoke about protecting our creative and copyright industries, we must apply that same protective rigor to our national health data.
“We need to ensure our data isn’t simply mined to find more treatments that drive new revenue opportunities for the pharmaceutical industry, rather than shifting towards a prevention-focused model,” she said.
“Imagine if Australia leveraged AI to build the next iteration of Medicare – a system that uses data and predictive technology not just to expand the volume of a “find and fix” model that pays to diagnose and treat, but one that actively prioritises predicting and preventing the need for care in the first place.
“Transitioning our technological focus from diagnosis, treatment, and management to true prevention is how we build a resilient, equitable, and sustainable healthcare system.”
Further reading
The PM referenced the Pope’s encyclical on AI, which addresses many of the determinants of health.
“…as with every major technological shift, AI tends to amplify the power of those who already possess economic resources, expertise and access to data. In light of the common good and the universal destination of goods, this raises serious concerns, since small but highly influential groups can shape information and consumption patterns, influence democratic processes and steer economic dynamics to their own advantage, undermining social justice and solidarity among peoples.
“For this reason, it is essential that the use of AI, especially when it touches on public goods and fundamental rights, be guided by clear criteria and effective oversight, grounded in participation and subsidiarity.
“Communities and intermediary organisations must not be reduced to passive recipients of decisions made elsewhere; they must be able to contribute to discernment and oversight. Moreover, ownership of data cannot be left solely in private hands but must be appropriately regulated.
“…the convergence of automation, robotics and AI is rapidly transforming the very structure of work. It is said that this will bring great improvements for everyone. In reality, however, the “new ways” of working are not necessarily better… the protection of employment opportunities and the irreplaceable role of the individual must remain the general rule.
“More than ever, in the age of AI and robotics, it is no longer possible to rely solely on the “invisible hand” of the market. Politics has the task of orientating economies and technologies to the common good, promoting dignified work, social inclusion and an equitable distribution of the benefits of innovation.
“A further risk, less visible but no less serious, is that of social control made possible by the massive collection of data and use of algorithmic systems.
“Even today, colonialism assumes new forms. It no longer dominates only bodies, but appropriates data, transforming personal lives into exploitable information. Entire regions, especially those marked by structural fragility and limited geopolitical relevance, are currently subjected to a new mindset of extraction: that of health data, epidemiological profiles, genetic maps and demographic information. These have become the new “rare earths” of power: vital data which, once aggregated and analysed, can be used to train predictive models, guide investment strategies, anticipate crises and, above all, determine who and what is deemed to matter.
“Those who control the health data of entire peoples – often collected under the pretext of aid, research or innovation – possess a structural leverage over the future, for they can shape needs and markets. They can also decide, before others, to whom medicines, investments and protections will be allocated.
“Here lies one of the most urgent moral challenges of our time: to ensure that shared knowledge becomes a true common good rather than an instrument of dominance. This requires restoring to individuals not only the data that describes them, but also the ability to decide how it is used, by whom and for whose benefit. Otherwise, the digital age will not be post-colonial, but colonial in another form.”
• Marie McInerney will cover The Royal Australian and New Zealand College of Radiologists conference on AI next week for the Croakey Conference News Service, with a preview article to be published on Monday, so stay tuned!
See Croakey’s archive of articles on AI and health
Add Comment