The Hidden Costs of Artificial Intelligence

By: Dr. Karen Becker President of the American Veterinary One Health Society

Diplomate of the American College of Veterinary Preventive Medicine

Smoke from Canadian wildfires is blanketing cities across the Mid‑Atlantic and breathing it may be equivalent to smoking a pack of cigarettes per day. Hospitals are bracing for spikes in respiratory illnesses while vulnerable residents are being urged to stay indoors, limit exertion, and monitor health symptoms hour by hour.

This crisis, immediate and personal, invites a broader question: should we continue accelerating the deployment of artificial intelligence (AI) without first examining its environmental costs? AI may indeed be the engine of the next great technological revolution, but the infrastructure required to power it is already straining the natural systems we depend on.

AI does not exist in the cloud. It exists in data centers—massive industrial facilities packed with high‑performance computers that consume staggering quantities of electricity and water. Electricity demand from U.S. data centers has already doubled and is projected to double again within the next several years. Similar patterns are emerging around the world in response to the adoption of AI.

When the grid cannot meet this new demand with clean energy, data centers turn to natural gas. The result can be higher emissions at the very moment nations are pursuing decarbonization goals.  AI and data center growth also require substantial water and land. Millions of gallons are required daily for cooling, often drawn from regions already facing water scarcity. The land needed to support data centers and associated power infrastructure continues to expand into agricultural and rural communities.

Beyond carbon emissions, water and land use, emerging science raises concerns over decreasing bee populations following nearby data‑center construction. While the research is still developing, studies have documented that electromagnetic fields from data centers can disrupt bee navigation, reduce foraging, and weaken colony health. As bees are essential pollinators for crops that feed millions, even minor disruptions can cascade across ecosystems and economies.

And yet the promise of what AI might achieve is difficult to ignore. In health care alone, AI is already detecting disease earlier, optimizing treatment plans, and easing administrative burdens that strain clinicians. Hospitals are adopting AI at extraordinary speed, seeing it as a necessary tool to alleviate staffing shortages, modernize outdated processes, and improve outcomes for patients who have long faced inequities in care.

AI can read imaging scans faster than radiologists, predict cardiac events before symptoms appear, personalize cancer therapies, and help distribute health resources more efficiently. In a nation where chronic illness is rising and access to care is uneven, AI has the potential to save lives, reduce costs, and widen access to medical expertise.

AI is a paradox. It undermines public health by using scarce resources, creating pollution, and harming eco-systems, but its existence has the power to improve public health.

The question is not whether AI is inherently good or bad. It is whether our current approach to its growth is sustainable, ethical, and aligned with long‑term public health and ecological stability.  If expansion continues unchecked, the negative will outweigh the benefits. Can we truly champion AI as a tool for human advancement while disregarding the environmental toll paid by the communities where its infrastructure is built?

We can choose a more balanced path.  Policymakers and industry leaders should require that new data centers source reliable clean energy—geothermal, nuclear, and grid‑integrated renewables—rather than defaulting to natural gas. Companies should prioritize energy‑efficient algorithms and invest in next‑generation cooling technologies that reduce water consumption. Regulators must update electromagnetic field standards to protect wildlife. And communities should have a voice in deciding where and how these facilities are built.

AI may help humanity flourish. But flourishing requires more than technological progress. It requires breathable air, drinkable water, food safety and security, healthy ecosystems, and a stable climate. These are not luxuries; they are preconditions for life.

As most Americans and Canadians experience the dangerous reality of climate‑driven air pollution, the stakes are clear. The benefits of AI cannot be evaluated only by its breakthroughs in medicine or economics. They must also be weighed against the environmental impacts borne by present and future generations.

The real question is not whether we can advance with AI, but whether we will demand that such advancement strengthens—rather than undermines—the foundations of our health, our environment, and our humanity.

---