While artificial intelligence consumes enormous amounts of electricity, researchers from the University of Waterloo and Georgia Institute of Technology found that AI's overall impact on global emissions is surprisingly small, and the technology could actually help accelerate the transition to clean energy. The findings challenge the prevailing narrative that AI is a climate threat, though they reveal important local complications around data centers. How Much Energy Does AI Actually Use? The numbers sound alarming at first. AI-related electricity use in the United States is comparable to the total energy consumption of Iceland, a country of roughly 370,000 people. A single ChatGPT request requires 10 times more electricity than a Google search, according to the International Energy Agency (IEA). Data centers that host AI technology consume vast amounts of energy, with the number of data centers increasing from 500,000 in 2012 to eight million today. Yet despite these staggering figures, the research suggests the climate impact at a national or global scale is manageable. "The increase in energy use is not going to be uniform. It's going to be felt more in the places where electricity is produced to power the data centers," explained Dr. Juan Moreno-Cruz, a professor in the Faculty of Environment at Waterloo and Canada Research Chair in Energy Transitions. "If you look at that energy from the local perspective, that's a big deal because some places could see double the amount of electricity output and emissions. But at a larger scale, AI's use of energy won't be noticeable". Where Is the Real Problem? The research reveals a critical distinction: while AI's global emissions footprint is smaller than feared, the local impact near data centers could be severe. Some regions hosting these facilities could experience a doubling of electricity output and emissions. This creates a geographic equity problem that deserves attention, even if the planet-wide climate math is more favorable than headlines suggest. Beyond electricity consumption, AI infrastructure creates additional environmental pressures. Data centers require advanced cooling systems that lead to excessive water consumption, while the short lifespan of graphics processing units (GPUs) and other high-performance computing components generates growing electronic waste. Manufacturing these components requires extraction of rare earth minerals, a process that depletes natural resources and contributes to environmental degradation. Can AI Actually Help Solve Climate Change? The more optimistic finding from the research is that AI could become a powerful tool for building greener technologies. "For people who believe that the use of AI will be a major problem for the climate and think we should avoid it, we're offering a different perspective," Moreno-Cruz said. "The effects on climate are not that significant, and we can use AI to develop green technologies or to improve existing ones". This potential is already being explored. AI systems can optimize energy supply networks by forecasting renewable generation and improving grid integration. In information networks, AI enables energy-efficient routing, resource allocation, and predictive maintenance. The technology can also help companies reduce their carbon footprint by optimizing energy use in buildings and data centers through predictive control systems and demand forecasting. How Can Organizations Use AI Responsibly? For companies claiming environmental commitments, the challenge is justifying AI use while maintaining credibility on sustainability. The solution isn't to deny AI's environmental costs but to deploy it strategically and transparently. - Choose efficient models: Use smaller or fine-tuned AI models for simple tasks like summaries and drafts, reserving heavier computational power for genuinely complex or high-value problems that justify the energy cost. - Select green vendors: Partner with AI providers that have strong renewable energy commitments, transparent data-center efficiency metrics, and services powered largely by renewables when possible. - Set internal guardrails: Establish clear guidelines for which tasks should and shouldn't use AI, discourage wasteful experimentation like thousands of near-duplicate prompts with no purpose, and batch tasks rather than running repeated single-use queries. - Compare to alternatives: Justify AI adoption by showing it replaces high-travel workflows, resource-intensive research and development, or manual repetitive reporting that would otherwise require more staff and office resources. - Maintain transparency: Be honest with employees about both the environmental costs and potential benefits of AI use, and involve them in identifying wasteful practices. What Do People Actually Think About Green AI? Public attitudes toward AI's environmental impact are shifting as awareness grows. A recent study surveyed 224 people about their willingness to support greener AI solutions. Over two-thirds of respondents expressed willingness to pay for more sustainable AI options, and higher environmental awareness was significantly associated with willingness to pay. This suggests that users care about the issue and may reward companies that demonstrate genuine commitment to reducing AI's carbon footprint. The challenge for employers and organizations is messaging. Many employees encouraged to embrace AI at work may be unaware of its environmental costs. As climate-conscious workers become more informed, they're likely to ask uncomfortable questions: "We're told to cut our carbon footprint, but now we're told to use AI for everything. How does that add up?" Credible answers require transparency about both costs and mitigation strategies. What's Next for AI and Climate? The research team plans to expand their analysis to other countries to better understand how AI adoption could influence energy use and emissions worldwide. Looking forward, three research directions are emerging: establishing low-carbon AI goals, improving energy efficiency in large language models, and developing AI-driven life-cycle management for hardware. The paradox of AI and climate is real, but it's not unsolvable. The technology uses enormous amounts of energy, yet its global emissions impact is manageable. The real opportunity lies in deploying AI strategically to accelerate the clean energy transition, while being honest about local impacts and committed to minimizing waste. For organizations serious about sustainability, that means treating AI not as a threat to environmental goals, but as a tool that must be wielded responsibly.
