Health and the environmentally responsible use of AI

We must ensure that as artificial intelligence becomes embedded in healthcare, it is evaluated not only in terms of clinical outcomes, but its broader environmental impact

In April this year, I attended the meeting ‘Artificial intelligence in healthcare: What clinicians in Ireland are saying’, organised by the RCPI. This event followed the recent publication of the AI for Care: The Artificial Intelligence (AI) Strategy for Healthcare in Ireland 2026–2030 by the Department of Health (DoH).

While it is clear that AI can play a vital role in enhancing patient care, I was struck by an important omission: The two-hour event failed to include a single mention of sustainability, emissions, climate change, or the wider environmental impacts of AI and data centres.

Similarly, the DoH’s AI strategy contains no reference to climate change or emissions, offering only limited mentions of sustainability. The DoH’s strategy is built on four main pillars: AI for clinical care; AI for operations; AI for research and innovation; and AI for public health. What is missing is a fifth, critical pillar: The environmentally responsible use of AI.

Data centres

AI is one of the world’s most resource-intensive technologies. Ireland is now home to one of the highest concentrations of data centres in the world. Data centres, the physical backbone of AI, already consume more than 20 per cent of Ireland’s electricity, 80 per cent of which still comes from fossil fuels.

A recently published report by the United Nations University Institute for Water, Environment and Health (UNU-INWEH) on the environmental cost of AI’s energy use highlights some stark statistics: In 2025, data centres, globally, consumed an estimated 448 TWh of electricity. Thus, if data centres were a country, it would rank 11th globally in terms of energy consumption. For perspective, 448 TWh of electricity is enough to supply the annual residential electricity needs of the entire population of Sub-Saharan Africa, 1.3 billion people, for over two and a half years. On current trajectories, data centre energy demand could double by 2030. This makes the environmental footprint of AI a serious concern that deserves attention alongside its clinical potential.

Back in Ireland, the rapid expansion of data centres is putting enormous strain on the energy grid. There are even data centres in Dublin with their own gas power plants. It is not surprising that the UNU-INWEH report highlights Ireland as a cautionary tale in a case study aptly named ‘Ireland: Small country, outsized data-centre load’. The pace of growth also raises questions about how data-centre demand interacts with the expansion of renewable energy and Ireland’s wider decarbonisation goals.

It is therefore concerning that, in 2026, such AI strategies and events still overlook these links. Doctors and healthcare workers have a profound responsibility to demonstrate leadership as we face the defining health crisis of the 21st Century: The triple crisis that is climate change, pollution, and biodiversity loss. While the energy use and consequent emissions may not be immediately visible to us in Ireland (beyond our rising energy bills), they cause tangible harm to both human and planetary health. The effects may be felt elsewhere, but they are no less real.

Data centres, the physical backbone of AI, already consume more than 20 per cent of Ireland’s electricity, 80 per cent
of which still comes from fossil fuels

Equity

This is also an equity issue. The environmental footprint and wider negative impacts of AI and data centres are disproportionately experienced by people and communities who benefit least from them. Locally, a recent analysis by TASC (Think-tank for Action on Social Change) found that nearly 90 per cent of Dublin’s data centres are in areas experiencing economic deprivation. Further scrutiny is needed to better understand how and where AI infrastructure affects people and communities, both in Ireland and globally.

That said, this is not an argument against using AI in healthcare. On the contrary, we should welcome strategies, frameworks, and open discussions. However, these technologies must be introduced in an environmentally responsible and equitable manner. We are still only beginning to understand the environmental footprint of Ireland’s healthcare sector. While we have some emissions data, we lack robust figures specifically regarding the emissions generated by AI and data centre usage.

Norwegian data offers an important warning. The Norwegian healthcare service estimates that nearly 12 per cent of their total emissions, over 180,000 tonnes of CO² equivalents, stem from information technology (IT) services. Norway is significantly ahead of Ireland in terms of healthcare IT integration, so this figure should prompt careful reflection. We face a choice to either allow emissions from IT and AI to rise uncontrollably, or to act now and shape their growth in a responsible manner.

If we are serious about evaluating healthcare innovation in terms of patient safety, quality, and equity, the environmental impact of AI must be a key pillar. Environmental accountability and sustainability should not be optional extras. They should be part of how we define responsible healthcare innovation.

The HSE has committed to becoming a net zero healthcare service by 2050, and that ambition should shape our approach to digital transformation. If healthcare is to lead on both patient outcomes and planetary health, environmental accountability must be a built‑in feature of all AI adoption, not an afterthought.

Role of healthcare workers

That said, doctors and healthcare workers must also extend their leadership beyond the clinic walls. While healthcare must do a lot more to reduce our own emissions, it is responsible for only ~5 per cent of global emissions. We must therefore get our house in order, but we must not stop there. Our message should be clear: AI can be very useful, but its rapid expansion is already increasing demand for energy, water, and raw materials, with harmful consequences for human health and the wider environment. If left unchecked, the growth of AI could deepen some of the very health and societal problems it purports to address.

While this piece has focused on AI’s energy use and emissions, its environmental impacts extend much further, from mineral extraction and its wider environmental and social harms, often in the Global South, to substantial water consumption, and poor management of electronic and hazardous waste. Across its lifecycle, AI and data centres carry consequences that ultimately affect human health.

Going forward, we must ensure that as AI continues to become embedded in healthcare, each new technology is evaluated not only in terms of clinical outcomes, but also its broader environmental footprint. Environmental accountability and sustainability should not be optional extras; they must be central to how we approach healthcare innovation. Only then can we introduce AI in a way that is equitable, holistic, and genuinely responsible.

It is the right thing to do.