AI was the topic during this Climate Week panel with Gavin McCormick (left) with WattTime and Climate Trace, and Tenika Versey Walker, head of Nvidia's Sustainable Futures program / Credit: Heather Clancy
L'Oreal, Philips, Salesforce and Tesla are just a few of the companies using AI to automate data gathering and decision making
Concern about the voracious energy appetite of large scale data centers used for artificial intelligence isn't dampening enthusiasm for its potential as a tool for fighting climate change.
AI humans identify climate risks and emissions reduction priorities with far more accuracy than previously possible, according to policymakers, sustainability professionals and consultants speaking out on the issue during Climate Week NYC. Global revenue from infrastructure, services and analytics related to AI could reach $15.7tr by 2030, forecasts research by PwC.
"We are standing on the precipice of a super cycle," said Mark Patel, senior partner at management consulting firm McKinsey, during the Climate Week NYC's opening session, discussing AI's ability to address a broad range of challenges. "We have an opportunity to turn this into something positive."
The island nation of Fiji, for example, used AI and freely available geospatial data to win funding for a system of natural seawalls to protect communities from rising sea levels in less than 12 months, said Patricia Scotland, secretary-general for the Commonwealth, an association that represents the interests of 56 member countries including Australia, Canada, India, Jamaica and Zambia.
"We are at the cusp of technological change that will enable us to write a whole new chapter," she said.
750-plus climate AI start-ups and counting
Chipmaker Nvidia, working closely with Amazon Web Services, Google and Microsoft on their AI technology infrastructure, supports more than 750 start-ups using AI to address agriculture, carbon capture, clean energy development, weather and climate risk modelling, waste management and other applications. The program was launched in 2023.
One of those ventures is Tomorrow.io, which uses space-based radar to provide predictive climate analytics and resilience data for customers such as Ford, Google, JetBlue, Microsoft and Uber.
"It helps you observe the world in ways you have never done before," said Cole Swain, vice president of strategy at the Boston-based start-up, during a climate and AI summit hosted by Salesforce and Nvidia. Tomorrow.io has raised almost $260m to date.
Practical applications for corporate sustainability team
Salesforce began using AI to gather data for emissions disclosure at least six years ago, and it turned that application into the foundation for Net Zero Cloud, which it introduced in 2020. AI far reduces the margin of error for those calculations and makes it much more auditable, which is a major consideration for companies preparing for the European Union's Corporate Sustainability Reporting Directive.
Salesforce published its corporate lobbying policies for AI in April, as concerns about the technology's environmental impact swelled. The company is carefully limiting the size and scope of the learning models for its AI technology, which helps reduce energy consumption, said Suzanne DiBianca, chief impact officer at Salesforce.
Right now, data centres account for about two per cent of the world's emissions, similar to the profile for aviation. The industry can limit growth of that footprint by choosing renewable energy to power AI models in regions where clean power is limited today and by being selective about the data sets, said Gavin McCormick, executive director of WattTime, and co-founding member of Climate Trace, an open source database made possible by AI.
"It could cause a climate crisis, but it doesn't have to," McCormick said.
Here are three other applications revealed during Climate Week:
- Healthcare tech company Philips uses AI to save a significant amount of time picking which suppliers to prioritise as part of its Scope 3 reduction strategy, an ongoing initiative. It did this by running thousands of data points through a model that screens for certain emissions factors and risks. "That helps us really to focus our attention and resource on things that really matter," said Robert Metzke, senior vice president and global head of sustainability at Philips.
- The world's biggest electric car company, Tesla, uses AI enabled by Climate Trace. Its inventory covers more than 352 million sources, including power plants, steel mills, ships, oil refineries, fertilizer applications, deforestation and wildfires. Tesla can use that information to switch supplier locations based on the energy used in a particular region.
- L'Oreal's research and innovation organisation uses AI to assess the emissions related to raw materials used in its fragrances and personal care products. The company sources more than 2,000 raw materials from 300 botanical sources. It screens them closely on a continuous basis for water stress, biodiversity impacts, agricultural practices and other climate and nature metrics, said Ezgi Barcenas, chief corporate responsibility officer at L'Oreal.
The common theme of these applications is automated decision-making, at a higher potential accuracy rate than otherwise possible. "Think about where you could use an assistant," and that is probably a task that would be well served with AI, said Salesforce's DiBianca.
This article first appeared at Trellis.
