Practitioners and researchers identify benefits and challenges of using AI to support sustainability, resilience, and social responsibility
May 31, 2026
- AI has the potential to help achieve sustainability goals by supporting decision making and solving complex resource allocation problems.
- Beneath the excitement are challenges: concerns about the misuse of AI by bad actors, over-reliance on the technology by well-intentioned users, and its effect on the environment.
- To overcome the environmental, economic, and organizational challenges posed by AI and ensure responsible use, humans must remain engaged.
Artificial intelligence can democratize research and open new avenues of discovery — but it also poses environmental, economic, and organizational challenges. Academics and practitioners gathered recently at Stanford Graduate School of Business to discuss the promise and the risks of AI technology, how it can be used to achieve sustainability goals, and how to make sure it is used responsibly.
“AI presents huge opportunities but also poses challenges in the way we do everything from research to teaching to running our day-to-day lives,” said Dan A. Iancu, Associate Professor of Operations, Information & Technology at Stanford GSB and organizer of the conference. “We have to learn how to best interact with AI as academics and also as society members.”
At the Sustainable and Responsible Operations in the Age of AI Conference, more than 100 attendees heard about how novel data sources and AI tools can support environmental sustainability, organizational resilience, and social responsibility. The event was part of the Stanford Initiative on Business and Environmental Sustainability Research Conference Series, hosted by Stanford Leadership Institute at the GSB and Stanford Doerr School of Sustainability.
Participants took a broad view of sustainability, encompassing not only the environment but also social resilience, economics, and governance, Iancu said: “It’s about taking a long-run view of objectives and impact, and not just looking at one metric but at multiple.”
Harnessing data for better sustainability decisions
One of the areas of promise for AI is in helping people make sense of huge volumes of data. AI enables researchers to quickly pull together diverse strands of research — something that can be helpful in solving problems that involve varying areas of research that have traditionally been treated in isolation.
“While AI will raise important challenges, it will also offer the potential to contribute to our understanding – and accelerate the solutions – for real sustainability programs, and we should try to harness that potential,” said Inês Azevedo, Professor of Energy Science Engineering at Stanford.
AI could also help accelerate the movement to integrate nature into economic decision making, said Gretchen Daily, Bing Professor of Environmental Science at Stanford, by making more accessible tools that can illuminate the benefits of nature, such as on mental health and many other core aspects of well-being.
“Humanity is deeply interdependent on nature,” Daily said. “Yet our decision making is mostly based on economic systems that were advanced a while back – a couple centuries ago – when nature seemed super abundant. How would we begin to understand the value of nature in urban systems?”
Another area where AI can help turn information into actionable insights is ocean conservation. Vast quantities of information about the earth’s oceans could, with the right analysis, be used to steer conservation efforts.
“The most exciting opportunity for AI in this space is to help us capitalize on that explosion of data,” said Jim Leape, William and Eva Price Senior Fellow at the Wood Institute for the Environment and Co-Director of the Center for Ocean Solutions at Stanford.
In healthcare, AI also has the potential to help improve patient care by surfacing the right information at the right time, said Mohsen Bayati, The Carl and Marilynn Thoma Professor of Operations, Information & Technology at Stanford. For example, he said, in cancer care, detailed research that could help clinicians with treatment decisions is often buried in long reports that the clinicians don’t have time to search through. AI can get relevant information to clinicians when they need it.
AI can also model how people make decisions — and it can aid that decision making by translating technical language into concepts everyone understands.
Irene Lo, Assistant Professor of Management Science & Engineering at Stanford, has been working with San Francisco Unified School District since 2019 to help design a system for assigning the district’s kindergarteners to elementary schools. AI tools helped her model how families are assigned to schools, improving the process to meet district and parent goals. AI tools have been helpful for translating between natural language and optimization language, unlocking tools for people who are not schooled in optimization.
Allocating resources and developing new technologies
Some of the most complex sustainability problems involve how to distribute resources fairly and efficiently — and there, too, AI holds promise.
Alexandre Jacquillat, Associate Professor at MIT Sloan School of Management, discussed his work on wildfire suppression, including helping the U.S. Forest Service develop tools to assist in triage decisions when multiple wildfires are burning at once. The difficulty of moving resources from one fire to another has to be taken into account, which may require policy changes.
“Environmental management matters. With optimization, a crew can have the impact of 1.8 crews,” Jacquillat said.
AI is also helping with the development and deployment of new technologies. Autonomous vehicles have the potential to improve safety, accessibility, and profits, but they also face challenges, including cost and opposition from labor unions and others who are concerned about replacing human workers. The vehicles also require complex routing systems that account not just for traffic and distance but also for hazards that could cause the vehicles to get stuck.
“Road safety is a significant issue worldwide,” said Daniel Freund, Associate Professor at MIT Sloan School of Management, who studies these vehicles and the issues involved in their deployment. “If self-driving cars can get us to have safer mobility, that is a significant advance in sustainability.”
Sounding cautionary notes
Alongside the excitement and the promise of AI, however, are risks. One issue raised by multiple speakers: concerns about the future of work, and how organizations can maintain institutional knowledge and avoid having AI cause massive layoffs.
There are also other concerns. For example, there is excitement about the potential for AI to monitor supply chains in real time. However, this is one of many areas in which bad actors could also use AI’s real-time analysis. For example, when AI has been used to predict wildlife migration patterns to protect migration corridors, poachers have also used the technology’s predictive capabilities to hunt the animals, said Sytske Wijnsma, Assistant Professor at the UC Berkeley Haas School of Business.
“The fact that noncompliant actors can now also use AI is something we’re currently not giving enough attention to,” Wijnsma said.
People may also use AI to take shortcuts, bypassing the necessary verification. In medicine, for example, doctors are using AI transcription tools to work more efficiently. However, Bayati said, “Potentially people can over-rely on these systems, and errors can harm patients.”
Another issue hanging over any discussion of advances in AI is the amount of electricity the technology will require. Azevedo noted that while there is a lot of uncertainty about how much electricity AI data centers will consume, some estimates suggest that within five years, the U.S. could see an increase in demand for electricity of over 20%. This raises questions about costs, emissions, and overall strain on the power grid.
“The country needs to be ready,” Azevedo said. “If there’s one thing that the country has not been good at, it’s building large scale infrastructure that enables sustainability (such as power plants, transmission lines, and other large infrastructure) as quickly as it needs to.”
Focus on responsible use
Speakers also explored how to use AI responsibly. They noted that human expertise is critical to responsible AI use: It still takes humans to understand all the facets of a problem, understand what different stakeholders need, and evaluate the results AI produces. And human creativity is still an essential ingredient.
“If you use AI to find better ideas, what are you gaining and what are you losing?” asked Léonard Boussioux, Assistant Professor at the University of Washington Foster School of Business.
In his experience using generative AI to augment the innovation process, for example, Boussioux said he found that AI could produce solutions that were, on average, better than humans’ solutions — but the most novel solutions came from humans.
“If you use AI, it will make you potentially more creative, but as a society, it decreases our collective pool of ideas,” Boussioux said.
And just as AI use leads to concerns about environmental and organizational sustainability, it also raises questions about the sustainability of research itself.
“Research is changing, whether we would like it to or not,” said Michael Wagner, Professor at the University of Washington Foster School of Business. He noted that AI has led to research advances in a number of fields, including high-level mathematics. But in many fields, AI can produce similar results to a junior researcher — which raises the question of whether there will be enough spots for graduate students. Likewise, the peer review system could become overwhelmed if researchers use AI to start producing a higher volume of papers.
Wagner also noted that what’s true today about AI may not be true tomorrow — AI is much more powerful today than it was a year ago.
“The ground is moving underneath our feet,” Wagner said.
Still, Wagner does not think researchers should fear being left behind. He noted that there are things AI cannot do yet: choose which problems are worth solving, evaluate novelty and significance, and teach the next generation of researchers. And AI may help expand research in unforeseen ways.
“AI can democratize research,” Wagner said.
Leave a Reply