Does AI Belong in Sustainability?

Would you use ChatGPT to help you write a climate action report when every query contributes to a growing environmental footprint? You wouldn’t be alone in hesitating. A recent poll by the University of Chicago found that 72% of Americans expressed concern about AI’s environmental impacts. That concern is translating into action: Data Center Watch reports that protesters blocked or delayed an estimated $98 billion in potential US projects between late March and June in 2025 alone. 

While AI devours resources at an alarming scale, it may present a powerful tool for driving transformative climate solutions. So, is it worth the trade-off? Can we harness AI for progress on climate goals without accelerating the crisis? Here, we explore the fundamental questions shaping perspectives on AI as a double-edged sword (Alnafrah 2025), exploring critical risks and potential opportunities to steer it toward a more sustainable future.

Why is AI an environmental concern? 

Artificial intelligence (AI) development accelerated dramatically in the early 2010s with the integration of NVIDIA GPUs. However, recent breakthroughs came with generative AI – technology that creates new content such as text, code, and images in response to user prompts. This has sent tech giants like Anthropic, OpenAI, Meta, xAI, and Google into a race to expand capacity and update the infrastructure AI relies on: data centers. These facilities consume significant amounts of resources. 

• Energy: Data centers are retrofitted with hundreds or thousands of servers, trained and ready to perform the computations that make up AI models. Their intense power demand raises concerns that clean energy goals could be sacrificed during rapid expansion. Energy consumption varies widely by task complexity – some estimates say a single ChatGPT query uses anywhere from 0.34 to over 20 watt-hours. For perspective, MIT calculated that generating a 5-second video on Meta’s Llama consumes the equivalent energy of running a microwave for over an hour. In 2024, US data centers alone used roughly the electricity needed to power the entirety of Thailand for a year (O’Donnell & Crownhart 2025).

• Water: Servers get hot while using this energy, requiring evaporative cooling using purified water from drinking water sources. Because the evaporated water is not currently collected for reuse, data centers raise serious concerns about water availability, particularly in already water-stressed regions. 

• Cost for communities: The strain on local resources is already being felt. Rising electricity prices, noise pollution, and water usage near data center hubs are fueling community concerns. Stephon T. Smith, an expert in banking, finance, and community development, warns that AI's rapid expansion risks sidelining equity and long-term planning—especially after the current administration attempted to terminate Greenhouse Gas Reduction Fund (GGRF) grant agreements, threatening safeguards for low-income neighborhoods. Without proactive intervention, Smith cautions, "we will build to solve the problem wrong, increasing energy bills and relying on outdated fossil fuel-intensive solutions, which puts us deeper in sustainability debt."

• Transparency and reporting: While some major AI companies have reported on data center impacts, methodologies are rarely comprehensive, and data has often been treated as a trade secret. Though Google has released its non-peer-reviewed methodologies for calculating impact, transparency remains limited across the industry.

What is the growing opinion about AI?

ChatGPT has been the most downloaded app for months and remains one of the most-visited websites globally. At Climate Week NYC 2025, AI dominated energy conversations and appears prominent at COP30, according to the UN. In general, many remain optimistic about AI’s innovation potential to protect people and the planet. 

Yet, opposition is growing. Many younger professionals whose educations, careers, and sense of selves developed during the technology boom are impatient with half-measures and crave AI strategy that is not only efficient but responsible, transparent, and designed for systemic change. “It can be done sustainably," says Jan Nowak, Procurement Sustainability Manager and community organizer at Xylem. “But that means real community input, actual environmental considerations…and a holistic, long-term approach–which few of these companies are taking.”

How can we make sure AI is used for good? 

Despite its challenges, AI's potential for environmental progress is real, if deployed collaboratively. In October, the Bezos Earth Fund—founded by Amazon’s Jeff Bezos—announced $30 million in investments for 15 organizations using AI to protect the planet, from reducing livestock methane and curbing illegal fishing to improving weather forecasting abilities and automating plant species identification. Google and global environmental nonprofit World Resources Institute (WRI) recently co-published a working paper highlighting how AI can scale action for nature, as long as it’s done in an “inclusive and responsible manner” (Gassert et al. 2025).  

AI can transform climate modeling, biodiversity tracking, and data-driven insights that are crucial for scaling climate solutions. But sustainability professionals need to shape AI development now, loudly and actively. Whether AI becomes a climate solution or an extractive technology depends on the boundaries we set today. Individuals, teams, and organizations must ask: how do we make sure it’s worth it? 

The answer starts with action: push for transparent policies, demand accountability, and come to the table ready to advocate for responsible AI deployment and strategy.

How have you shaped the trajectory of AI in your organization? Start by reaching out to a colleague in another team or sector to compare notes. I’m happy to be a sounding board – feel free to connect. 

Note: Opinions expressed are my own and those of as select peers, reflecting the state of AI as of November 2025. 

Works Cited

Alnafrah, I. (2025). The two tales of AI: A global assessment of the environmental impacts of artificial intelligence from a multidimensional policy perspective. Journal of Environmental Management, 392, 126813. 

Gassert, F., Gawel, A., Harfoot, M., Mayer, A., Singhal, K., Stolle, F., & Vary, L. (2025, November 4). AI for Nature: How AI Can Democratize and Scale Action On Nature. World Resources Institute.

O’Donnell, J., & Crownhart, C. (2025, May 20). We did the math on AI’s energy footprint.