AI + Healthcare ≠ Climate Progress | The Case for Healthcare AI Sustainability

Adam Peddicord
5 days ago
7 min read

A Gross Oversight

I'd completely overlooked it and need to acknowledge I'm part of the problem.

This bothered me because I fancied myself a leader in responsible AI implementation and sensible governance.

My head-down, hyper-focused thinking blinded me to what should have been an obvious cost and trade-off pursuing AI so aggressively in healthcare.

Patient outcomes and experience, practitioner adoption and transformation, payer fluidity, policy development all were top of mind as I've been aggressively pursuing the next evolution of patient care.

It wasn't until I attended the Climate Change and Healthcare panel at Experience Momentum last week that the proverbial anvil landed on my head.

Healthcare in the US currently accounts for 8.5% of our national carbon emissions.

No one has yet reliably quantified how much AI's rapid adoption in healthcare will add to that footprint. That absence is part of the problem.

The data center, power grid, and operational cost of AI dominate a lot of discussions today, and rightfully so. But in our narrower lanes of expertise we may be overlooking the true cost of advancement, which must be a line item in our assessment of how to proceed with effective AI implementation.

What the data says.

The Climate Crisis's Impact on Healthcare

Living in the Pacific Northwest, I've watched hikes require fewer layers, hazy skies replace fog, and air conditioning go from an oddity to a necessity. The data backs me up.

Research, trends, and data analysis all clearly show climate change is having a negative effect on the health of our population. Global temperatures have risen roughly 1.1 to 1.2 C since the late 1800s (IPCC AR6) driving extreme heat, degraded air quality, shifting infectious disease, and more severe weather, all of which raise the human and financial cost of care.

Extreme heat doesn't just increase the likelihood of heat exhaustion, it acts as a chronic-disease multiplier, pushing existing heart, kidney, and metabolic conditions over the edge. Currently an estimated 546,000 heat-related deaths occur every year globally.

Air quality is degrading across three fronts: ground-level ozone, fine particulates, and longer pollen seasons, driving respiratory and cardiovascular disease, with an estimated 100,000 to 200,000 US deaths a year tied to air pollution.

Infectious disease is expanding its range. Warming lengthens vector seasons and moves Lyme, West Nile, and dengue north. The US logged 6,543 locally acquired dengue cases in 2024, more than the prior ten years combined.

Extreme weather hits patients and the system at once. Floods, hurricanes, and heat drive injury, displacement, and a heavy mental-health toll of elevated PTSD, depression, and anxiety, while surging demand at the very moment they knock out capacity, forcing earlier discharges and rationing as rising energy costs strain already-thin margins.

With extreme-heat days increasing around the US (over the last decade Seattle's average number of days at 90+ degrees has gone from 3.5 to 7.5), wildfires occurring more consistently, and a myriad of other climate-driven challenges emerging, it is safe to project climate's impact on stressing the healthcare system will only continue to increase. We are adopting AI to improve health, yet if it worsens climate it feeds the very trend making people sicker.

AI's Role in Climate Challenges

Not all AI is created equal when it comes to climate impact. Therein lies both the challenge and the opportunity. A simple tool automating a routine task barely registers.

But depending on the model, a heavily adopted hospital system leaning on a large LLM for administrative work can carry the daily emissions of a car driven around 30 miles, every day.

AI is also a major new load on the grid. Data center electricity is set to roughly double by 2030, with US demand up about 130%, and NERC is flagging tightening reserve margins. Hospitals sit right in the blast radius. They run about 2.5 times the energy per square foot of a typical commercial building, with electricity alone accounting for close to 70% of their utility costs, which leaves them exposed to both rising energy costs and reliability risk.

Why Healthcare + AI ≠ Climate Change Progress

Intuitively, the two look like a terrible marriage for the planet. Climate change is already driving up the burden of illness, and AI adds real and projected stress to the environment. But the data reveals it is not so black and white.

First, peer-reviewed work shows AI can also cut healthcare emissions, mainly by displacing travel and in-person visits, with autonomous AI screening reaching reductions as high as 80% in some settings.

The footprint is implementation-dependent and largely unmeasured. Presently there are no credible forecasts that isolate AI-in-healthcare's emissions, and current reporting frameworks don't account for them. So the net impact turns on design choices, not the technology itself.

“Right-sizing” is the highest-leverage move for balancing AI's impacts. Task-specific and small models are orders of magnitude more efficient than large general-purpose LLMs for the same job, so the first question is whether you need a big model, or AI at all.

Finally, I was unable to uncover any evidence where environmental cost is an existing line item in AI scoping.

Every AI business case asks:

✓ What's the ROI?

✓ What's implementation cost?

✓ What's adoption risk?

Almost none ask:

❓ What's the environmental cost?

After spending years overseeing AI operations and helping organizations operationalize AI, I've learned that success rarely comes down to the model itself. It comes down to what leaders decide to measure.

The Opportunity for Healthcare Leaders

There's immense pressure on healthcare organizations to adopt AI. Some is projected to solve a real need aligned with better patient outcomes. Much of it is following the hype without proper diligence done to determine the true effectiveness of the implementation.

Stepping up to accept the realities of our evolving planet and understanding where the benefits (and costs) of AI sit within your organization are essential to uncovering the best path forward.

I would also argue staking your claim as an AI-forward and environmentally conscious healthcare organization differentiates you in the most positive light. You get it's about better patient outcomes. You are on top of all the components that drive it so much so that you've adopted a framework that balances AI progress with its costs. You're not playing the game, you own the game.

What Leaders Should Do

Be transparent about the data and trends with regards to AI in Healthcare's direct impact on the climate crisis. A lot remains unknown. While intuition suggests AI adoption won't be positive for the climate footprint, the reality is there's not a lot of concrete data for healthcare specifically in this area.
Adopt a climate/environmental footprint policy and incorporate it into your cost-benefit analysis when exploring AI adoption. “Nutrition labels” and impact calculators are being constructed in varying forms. While no single representation seems to have become “the standard” for all to follow, the point is to have one and use it. You may always refine it as you go.
Remain open-minded on where the wins may come from. It may be as simple as the product you're choosing. A simple switch there to something that's just as effective for the patient but less stressful to the environment may be the offset you're looking for while AI's cost remains high.
Use the five E's of AI for decision-making. It's a simple framework that forces you and your team to think through the proposed solution and what its real impact may be.
- Efficiency: Does this actually save meaningful effort or time, or does it just feel faster?
- Effectiveness: Does this actually improve the outcome I am looking for, and how will I verify the result is accurate?
- Ethics: Is this an appropriate use of AI, given the context and the people involved?
- Equity: Who benefits from this use of AI and who might be left out or harmed?
- Environment: Does the potential environmental cost of this AI use match the value it offers?

Final Thought

Today I awoke to another heat advisory warning on my phone. There is also concern that winds will begin to blow smoke from wildfires into our area impacting our air quality for the week at minimum.

We live in a home where we thought air conditioning and air purifiers wouldn't be necessary. Yet one of our kids develops a constant cough anytime our summer hits this season.

All our windows are shut, blinds drawn, and I'll be installing an AC unit later today to try and keep the home livable for our family. Others aren't as fortunate as us.

There will be heat-related casualties this week in the Emerald City, and elsewhere. You don't need an AI algorithm to forecast that.

My belief is that within a few years, environmental impact will become another required operating metric.

Organizations that begin building that capability now will make better investment decisions long before regulators require them to.

If your organization is evaluating AI initiatives, I help healthcare and SaaS leadership teams pressure-test AI initiatives before they become expensive operational mistakes.

I help healthcare and SaaS organizations improve Customer Success, Revenue Operations, and AI Operations by designing scalable systems that reduce operational friction, improve customer outcomes, and allow teams to perform at a high level without burning people out.

If that sounds familiar, I'd be happy to compare notes and explore what's getting in your team's way.

Book 15 min →

Have you ever seen environmental impact included in an AI business case?

Cheers,

Adam Peddicord

Customer Success by Design

Sources

Abramoff, M. D., et al. (2023). Potential reduction in healthcare carbon footprint by autonomous artificial intelligence. npj Digital Medicine, 6, 65. Read →
Beaglehole, B., Mulder, R. T., Frampton, C. M., et al. (2018). Psychological distress and psychiatric disorder after natural disasters: systematic review and meta-analysis. The British Journal of Psychiatry, 213(6), 716–722. Read →
Eckelman, M. J., Huang, K., Lagasse, R., Senay, E., Dubrow, R., & Sherman, J. D. (2020). Health Care Pollution And Public Health Damage In The United States: An Update. Health Affairs, 39(12), 2071–2079. Read →
Goodkind, A. L., Tessum, C. W., Coggins, J. S., Hill, J. D., & Marshall, J. D. (2019). Fine-scale damage estimates of particulate matter air pollution. Proceedings of the National Academy of Sciences, 116(18), 8775–8780. Read →
Intergovernmental Panel on Climate Change (2021). Climate Change 2021: The Physical Science Basis. Working Group I Contribution to the Sixth Assessment Report (AR6). Cambridge University Press. Read →
International Energy Agency (2025). Energy and AI. IEA, Paris. Read →
Lawrence Berkeley National Laboratory (2024). 2024 United States Data Center Energy Usage Report (Shehabi, A., et al.). U.S. Department of Energy. Read →
Luccioni, S., Jernite, Y., & Strubell, E. (2024). Power Hungry Processing: Watts Driving the Cost of AI Deployment? Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency (FAccT). Read →
North American Electric Reliability Corporation (2025). 2025 Long-Term Reliability Assessment. NERC. Read →
Ramachandran, et al. (2025). Sustainably Advancing Health AI (SAHAI). NEJM Catalyst Innovations in Care Delivery, 6(10).
Romanello, M., et al. (2025). The 2025 Report of the Lancet Countdown on Health and Climate Change. The Lancet. Read →
U.S. Centers for Disease Control and Prevention (2025). Dengue in the United States: 2024 Case Data. CDC. Read →
U.S. Energy Information Administration (2022). Commercial Buildings Energy Consumption Survey (CBECS): Health Care Buildings. U.S. Department of Energy. Read →
U.S. Environmental Protection Agency (2024). Climate Change Indicators: Heat Waves and Heat-Related Deaths. U.S. EPA. Read →