April 21, 2026
When AI meets climate action
In honor of Earth Day, a graduate student shares her five-part series of 'AI What-If Scenarios for the Planet'
UNIVERSITY PARK, Pa. — “Our Power, Our Planet.” The 2026 Earth Day theme serves as a reminder that environmental progress, such as emissions reduction, is sustained by the daily actions of communities, workers and families protecting where they live and work. Technology like artificial intelligence (AI) is a powerful tool for this progress, but its success depends on global cooperation and careful regulatory oversight.
“AI is both part of the problem and part of the solution, and we need to be honest about that duality,” said Shikha Soneji, who will graduate next month with a Ph.D. in informatics from the Penn State College of Information Sciences and Technology (IST). “The very tools we are proposing to fight climate change have their own growing carbon footprint.”
Soneji’s doctoral training outlining the research methods, the grounding in information science and policy and the technical depth in natural language processing and machine learning has shaped how she thinks about the human impact of technology. She recently co-created* the “AI What-If” series, which explores how AI might enhance climate policy through data-driven simulations that evaluate large-scale sustainability strategies.
“We wanted to make complex AI-and-climate scenarios accessible to a general audience — to get people thinking about real consequences before they become normalized,” Soneji said. “The goal was to move the conversation from abstract ‘what could happen’ to ‘this is already happening.’”
Here are her five scenarios to explore how AI can enhance climate policy:
1. What if half the world never commuted to work again?
If 50% of jobs became permanently remote, almost 1 billion tons of CO₂ could be saved every year by reducing commuting and office energy use. However, those gains might be offset by data center energy demands as AI and cloud systems expand. This is exactly the kind of research needed to counter digital progress with environmental impacts, according to Soneji.
Remote work shows us how lifestyle shifts can cut emissions dramatically, but it also highlights the hidden costs of our digital world. AI can help balance the equation by optimizing cloud workloads, powering smart homes and reducing waste in virtual collaboration. But policy, infrastructure, and sustainable design must keep up.
2. What if every watt of power came from the sun, wind, and/or water?
If the world transitioned to 100% renewable energy by 2040, the shift could potentially eliminate 70% of power sector emissions, create millions of green jobs, and save trillions in climate and health costs.
AI is already doing some heavy lifting by predicting solar and wind patterns, managing smart grids and preventing system failures before they happen. To work, investments in infrastructure and innovation are needed. There is no dearth of renewable resources, and there is endless potential to harness them.
3. What if every country kept its climate promise?
In 2015, the world said "yes" to the Paris Agreement to limit warming to 1.5°C and to cut emissions by 43% by 2030. But in 2026, most countries are nowhere close. Countries can opt in, step out, change course.
AI can help by tracking emissions from space, simulating policies, and optimizing how we power cities, move goods, and grow food. But beyond the tech is the willingness of countries to align and commit to real change.
4. What if AI made us use less energy, not more?
AI is often blamed for driving up energy demand. But automation and intelligent systems can also help save energy.
Already, AI is helping buildings adjust HVAC systems in real time, helping factories use predictive maintenance to avoid waste and helping smart grids balance demand before blackouts happen. But the very tech that helps save energy also uses a lot of it. It will come down to how systems are designed, deployed, and regulated.
5. What if every new building were net-zero by 2030?
Buildings are responsible for 39% of global CO₂ emissions. Imagine a future where every new building produces as much energy as it uses, where smart HVAC, efficient materials and AI-powered retrofits are standard.
AI is already helping by designing low-carbon buildings, optimizing energy use in real time, and detecting inefficiencies before they show up in consumers’ utility bills. But upfront costs are high, regulations are messy and retrofitting older buildings is challenging. Real climate progress will come not just from smarter tech but from smarter standards as well.
Bottom line:
“Technology alone is not enough — progress requires a framework for integrating AI-driven climate analysis into real decision-making,” Soneji said. “Governments have a critical role in ensuring that AI is deployed effectively to accelerate the transition equitably and sustainably. Collaboration is needed among environmental scientists, data scientists and policymakers.”
*Shikha Soneji’s Ph.D. in informatics is focused on natural language processing and machine learning. She co-created the AI What-If series with Sameer Neve, who holds a Ph.D. in environmental engineering and has a deep knowledge of sustainability. Read their paper, "Emissions Impossible: Five AI ‘What If’ Scenarios for Emission Reduction” at IEEE Explore.
Image credit: www.svma.co.uk