“We need to start asking people and corporations, how can you drive an impact in a way that makes it about more than just you,” said Lucas Joppa, Microsoft’s chief environmental officer, at a Monday talk about leveraging artificial intelligence (AI) and machine learning to mitigate the impact of human activities on the planet.
The discussion, the third installment in the “AI for Good” seminar series, also featured assistant computer science professor Stefano Ermon and was moderated by Gretchen C. Daily, an environmental science professor and director of Stanford’s Center for Conservation.
Ermon began by highlighting recent progress in AI research, including in areas like computer vision, speech recognition and game playing.
“We need to think about how we can use these techniques so that they end up benefiting as many people as possible,” Ermon said.
He added that it is important to have representative data, and proposed developing models that would provide insight into issues like infrastructure quality, food insecurity and poverty.
“This is just the first layer,” Ermon told the audience. “We can build a whole stack of solutions on top of these.”
Joppa pressed the importance of deploying technology to maximize humans’ benefits on Earth’s systems.
“I have focused my academic time thinking about what goals we should be setting as an international society and how we might go about achieving them,” Joppa said. “More importantly, I’ve been thinking about how good of a job we’re doing at completing them already.”
After writing a memo entitled “AI for Earth” that made recommendations about how Microsoft should deploy its technology after investing in AI research for 30 years, Joppa spearheaded a team within Microsoft to tackle this question.
He described principles of his work, including recognizing that money should not be a barrier for companies who want to apply a machine-learning approach to environmental challenges and acknowledging that machine-learning talent is in short supply.
“If we don’t put the talent we have into play, we are far from where we can be,” Joppa said. “Building up engineering and machine-learning teams within companies can help.”
Asked about how they personally were working on furthering the application of AI to protecting the environment, Ermon and Joppa both stressed the need for gathering data.
“We live in the most narcissistic definition of an information age,” Joppa said. “We have so much information about ourselves and so little information about the planet we call home.
Ermon described the work that he has been doing to remedy this problem.
“We’ve been using social media, and played around with the idea of using data from Twitter,” he said.
Joppa described actions by Microsoft to mitigate its effects on the environment, such as supporting organizations using technology to mitigate humanity’s environmental impact. Microsoft has also committed to being carbon negative by 2030.
“It was depressing to see how many people thought that Microsoft’s commitment was so ambitious when really all we did was what the IPCC says everybody needs to do,” Joppa said, referring to the International Panel on Climate Change.
Liana Keesing ’23 praised the speakers for their dedication to using technology and data to benefit the environment.
“What really interested me was the discussion of how to optimize our data so that we’re able to create programs for environmental protection using cheaply available data that already exists out there,” Keesing said. “There is so much that we can take advantage of to make an impact.”
Contact Sarina Deb at sdeb7 ‘at’ stanford.edu.