In the European Union, large companies and most publicly traded firms will be required to publish updates on the environmental and social risks they face, with those reports due beginning in 2025.
Across the pond, the Securities and Exchange Commission earlier this month announced new rules that would require corporations to divulge to investors their greenhouse-gas emissions. “It’ll make it real for a lot of chief financial officers,” says John Mennel, managing director at Deloitte.
With disclosure requirements emerging in nations ranging from India to China, the demand for tools that help firms track environmental, social, and governance factors are expected to rise. Many of those tools are going to lean on artificial intelligence to help multinational companies remain compliant—but also transform their businesses along the way.
“What you see is about 40% of the world GDP now requiring not just climate disclosure, but full sustainability—the E, the S, and the G based on what’s called this sort of double materiality task, which is not just effects on the company, but the company’s effects on the world,” says Tim Mohin, partner and director at Boston Consulting Group.
“I think it’s really great for businesses,” says Meera Clark, an early stage investor at Redpoint Ventures. “Now that they have the visibility, they have this benchmark that they’re building towards as it relates to really setting up the reporting infrastructure that they will hopefully be able to rely on for the next 5, 10, or 20 years.”
AI is already helping companies track their ESG objectives in a wide variety of ways, with some of the most popular use cases ranging from machine learning to improve the accuracy of ESG metrics and address disclosure gaps, to using AI-powered satellite technologies to assess environmental risks, and predictive modeling to calculate greenhouse gas emissions.
“There’s just a whole range of areas where AI will play a meaningful role,” says Matt Slovik, head of global sustainable finance at Morgan Stanley.
But, Slovik adds, companies should proceed with some caution. “Is this a problem that AI can help solve?,” Slovik asks. “And if so, what does that solution look like and what does it mean within the context of your organization, your cost structure, and your other goals to ultimately get to the right decision.”
At Redpoint Ventures, Clark says the firm spends a majority of their focus on software and data infrastructure companies, looking for mass market opportunities and a key reason why a startup didn’t exist previously but should today.
“The regulatory environment continues to evolve,” says Clark, which in her view justifies Redpoint Ventures’ led investment in a $13 million Series A for AI sustainability platform Greenplaces. “There’s a very clear need for businesses to be able to more effectively report on this data.”
One question that’s emerging amid the rise of demand for AI and generative AI tools is the energy use needs for such computation. “Data centers continue to consume an outsize portion of energy and unless that energy is sort of sourced renewable or there is some way to mitigate its actual consumption, we’re going to have a bigger and negative side to the story,” warns Mohin.
Susannah Shattuck, head of product at AI governance software provider Credo AI, says that if an organization has set a target to achieve carbon zero by 2050, they need to make risk-informed decisions and be “aware of the fact that these large language models can have a tremendous carbon footprint and therefore be making decisions about deploying them really in the use cases that have the potential to have the greatest impact on my business.”
Large language models that can contain hallucinations, bias, or subject an organization to adversarial attacks can cause a model to lead them in the wrong direction, resulting in new governance risks when companies rely on these tools.
“An organization that wants to make use of that technology safely, needs to ensure that they’re the proper guardrails in place to protect against those possible risks and negative outcomes,” says Shattuck.
At Deloitte, Mennel says AI tools for ESG not only helps companies remain compliant with new standards, it actually can transform them. An agricultural company can use AI to track the environmental footprint of a new, lower carbon alternative source of protein, as an example, and market those claims to consumers. “With the data that I can produce, where are there the opportunities to create fundamentally new products or new businesses that generate value,” asks Mennel.
Canada-based Geotab has used AI for more than a decade to help Coins2Day 500 firms and the public sector manage their fleet, offering data intelligence to make more informed decisions about the efficiency, safety, and sustainability of the vehicles they use. “Sometimes there’s a strong overlap between some of the sustainability decisions and efficiencies,” says Neil Cawse, CEO of Geotab.
The most sustainable solution is a simple one: reduce vehicle idling. From there, Cawse says the public and private sectors can ascertain the range of their routes, and make sure the vehicle type matches the requirements of the route. But one common mistake is that some companies switch too aggressively to electric vehicles, a costly error if some vehicles in the fleet are sidelined because they cannot complete a journey.
“Good decisions are driven by good data,” adds Cawse. “Let the AI figure out what it is that you need to focus on first.”
In February, Geotab unveiled a generative AI analytics assistant called Geotab Ace, which taps into billions of data points daily including predictive safety analytics, predictive maintenance, trip data, EV statistics, and GPS tracking to answer questions from customers.
C3 AI, meanwhile, sells AI-enabled software to sustainability teams to help them gather, manage, and analyze data, identify risks, and execute plans to meet their ESG goals. “The demand for these ESG applications is about to go through the roof,” says Tom Siebel, chairman and CEO of C3 AI.
Siebel questions if ESG disclosure requirements will make a meaningful difference, saying such standards will be enormously expensive for firms to report. Firms only need to publish the data to comply with regulations emerging from Europe, the U.S., and other markets—there’s no real requirement for action.
C3 AI’s hope is that the tools it offers will result in action, proving a reduction in firms’ energy footprint will reduce costs and ultimately be better for customers, shareholders, and the planet.
“We will allow them to plan and mitigate measures to reduce their carbon footprint and reduce their energy footprint to meet the goals that they’ve set,” says Siebel.
