AI-driven emissions + sustainability reporting: a love story

This originally appeared in the July 2024 ENGAGE newsletter. Subscribe to have articles like this and other sustainability communication content delivered directly to your inbox each month.

Artificial intelligence is creating quite an uncomfortable moment for corporate sustainability and broader decarbonization efforts. 

A growing number of companies are finding their net-zero goals under threat by the burgeoning AI economy’s gluttonous energy appetite. While businesses certainly can’t be blamed for wanting a piece of that tasty $15.7 trillion pie AI will represent by 2030, this potentially is coming at a high cost for climate action. 

I say “potentially” because when it comes to AI’s ultimate impacts on decarbonization efforts, the jury is still out. 

Some studies claim that AI technology could grow to consume the same amount of electricity used by the entire country of Ireland. And this is happening too fast to be balanced by offsets and investments in renewables and energy efficiency. Conversely, a Boston Consulting Group study claims that AI has the potential to help mitigate between five and ten percent of global GHG emissions by 2030 — the equivalent of roughly the total annual emissions of the EU. Mitigation aside, AI also shows promise for advancing climate adaptation by predicting extreme weather, improving energy efficiency and optimizing resource use, among others. 

Assuming that AI doesn’t evolve into some self aware being that destroys civilization as we know it, today’s companies and the humans that run them are going to need to figure out a future that allows us to leverage AI without exacerbating the climate crisis. And this involves how companies communicate around AI sustainability. 

Recently, a client asked for advice on how to address AI-driven emissions increases in their upcoming sustainability report — and how to position this in their broader sustainability narrative. With this becoming an increasingly common conundrum for companies grappling with AI sustainability issues, let’s take a closer look.

Transparency is the best policy

As a general rule, companies should be open about why their emissions have increased — and the steps being taken to address it. Omitting this information from a sustainability report will only spark questions from curious stakeholders. 

Due to AI’s megatrend status, this isn’t something companies can easily sweep under the rug. This is doubly true for hyperscale cloud service providers like Google, Microsoft and Amazon that offer computing power and storage to organizations and individuals around the world. These hyperscalers own the data centers where the world’s growing number of AI algorithms live — and this is putting intense pressure on their Scope 1 and 2 emissions due to increased electricity demands. And Scope 3 emissions also are shooting up due to new construction and the addition of new graphics processing units in the supply chain. Microsoft Azure, Google Cloud Platform and Amazon Web Services alone account for more than 60 percent of the cloud market — which is putting enormous pressure on these leading companies to procure offsets, renewable energy and energy efficiency measures faster than is currently possible. 

“For some technology-first companies, AI will represent an increase in emissions that needs to be monitored and managed,” Matt Sekol, Sustainability Global Black Belt at Microsoft and Author of ESG Mindset, recently told me in an email. We also dove deeper into AI sustainability in a recent episode of The Sustainability Communicator. 

In its 2024 environmental sustainability report, Microsoft — which has pledged to be carbon negative by 2030 — disclosed that its emissions grew by 29 percent since 2020 due to the construction of more data centers designed and optimized to support AI workloads. In the report, Microsoft discusses the positives of AI from a high level before mentioning its sustainability drawbacks: “This year, technologies like AI brought renewed promise of the role innovation can play in accelerating progress… Amid this optimism, we face the realities of the complexity of the challenge.”

Likewise, Google’s emissions shot up 13 percent in 2023 driven in large part by AI. While this is quite startling for a company that has long prided itself on being on the forefront of climate action, Google was smart to get out in front of the problem by leading its 2024 Environmental Report with a section called ‘AI Sustainability.’ 

“We know that scaling AI and using it to accelerate climate action is just as crucial as addressing the environmental impact associated with it,” writes Google CSO Kate Brandt in the report’s executive letter. 

The takeaway — if a company builds, owns or operates data centers, AI-related emissions increases should be addressed in their sustainability report in a balanced way mentioning the risks and opportunities. And this should be accompanied with meaningful strategies for addressing AI’s environmental impacts in the long term. 

AI and sustainability reporting for everyone else  

With most companies not being Google, Microsoft or Amazon, AI sustainability remains a material challenge for those working to address indirect emissions in their broader value chains. Cloud emissions and cloud AI workloads, for example, would fall under Scope 3 for many companies. 

“As with any new technology or even product offering, emissions will go up because it is a new activity,” Sekol said. “For the company, it represents an intangible, like IP, but uses tangible resources such as in a cloud datacenter somewhere.”

There are governance and management principles a company can adopt to ensure that they are managing their models efficiently. Still, there’s a high dependence on data centers to secure offsets and renewable energy for those power-hungry AI models. And yet together, they can work to design and deploy AI to address climate challenges.

“When enterprises design and deploy AI sustainably, they can unlock the opportunity to help solve environmental and social challenges simultaneously while driving economic growth,” datacenter provider Equinix wrote in a blog post earlier this year. “Developments are underway to leverage AI to reduce operational carbon emissions and accelerate climate action programs.”

Ultimately, companies must clearly communicate plans for addressing the sustainability impacts of AI in their value chain. If a company lacks one during this year’s reporting cycle, they ought to disclose that they’re working on it (and mean it). And as I’ve written in GreenBiz (now Trellis), your sustainability report isn’t enough to effectively communicate your sustainability story. Remember to find ways to weave your AI sustainability efforts into other mediums like your blog, social media, internal newsletters and however else you reach your stakeholders.

Moving forward, AI sustainability is going to be a growing reporting and communication challenge for companies around the globe.

How is your company grappling with AI Sustainability in your strategy and communication efforts?

We’d love to hear it. And as always, feel free to reach out with any questions on this or any other sustainability communication topic.