Can businesses incorporate energy-intensive generative AI without sabotaging their sustainability goals?
A recent study estimates that global AI demand could cause data centers to use 1.1 trillion to 1.7 trillion gallons of fresh water by 2027. Photo by Getty Images on Unsplash
By using generative AI in your business, you are accelerating the climate crisis.
But that might be the right thing to do.
“There are a couple of ways to think about this topic. One is where AI is going to be an enabler for sustainability outcomes and imperatives. And the other side is that the rush to AI is going to create challenges around sustainability,” Rika Nakazawa told Digital Journal.
Nakazawa is the current Group Vice President, Connected Industry and Head of Sustainability for Americas, at NTT Data, a multinational IT services and consulting company. Nakazawa previously held roles at NVIDIA and Accenture. She is also the Co-Founder of Strides AI.
Her comments echo a sharpening view of the dichotomy of AI’s promise and potential perils. It can help optimize renewable energy production, reduce our reliance on fossil fuels, and enhance our climate modeling capabilities. But it can also create drains on our resources that can lead to water shortages, sudden brownouts, and spikes in our collective carbon footprint.
Gen AI uses massive amounts of energy and water
It’s important to be clear about the resource intensity of generative AI (gen AI).
In 2022, the world’s data centres used 460 terawatt hours of electricity. The International Energy Agency (IEA) predicts this will more than double to 1,000 terawatt hours by 2026. This is effectively the energy consumption of Japan (population: ~125 million). This spike will be driven heavily by generative AI. To wit: According to the IEA, a single Google search takes 0.3 watt-hours of electricity, while a ChatGPT request takes 2.9 watt-hours.
But AI doesn’t just drive spikes in energy usage. It also consumes water, which is used to cool down data centres. The Atlantic recently reported on a UC Riverside study that estimated that global AI demand could cause data centers to use 1.1 trillion to 1.7 trillion gallons of fresh water by 2027. This happens against the backdrop of a world where, “nearly two-thirds of our world’s population experiences severe water shortages for at least one month a year, and by 2030, this gap is predicted to become much worse, with almost half of the world’s population facing severe water stress.”
However, it’s unclear if business leaders have seriously considered the impacts of widespread gen AI usage on their sustainability goals.
“Gen AI and sustainability will be very pressing for companies that can be hurt by high energy costs. But a lot of that is going to be borne by the data centres because many companies are not going to be running AI applications on-premises or in their facilities,” said Nakazawa.
“So, they’re not going to see the electricity bill. They’re going to see the cost of hosting and cloud services go up because of energy intensity. But they may end up not caring about sustainability as much as they care about it as a secondary aspect of the cost of energy.”
The data suggests she’s right.
A 2023 PwC Emerging Tech Survey indicated that only 22% of business leaders cited sustainability impact as a top issue in gen AI deployment. An Ernst & Young survey indicated that 71% of UK CEOs agreed that activist shareholders are more concerned with their company meeting quarterly earnings targets than its performance against long-term sustainability metrics.
Related content: Does sustainability matter for businesses in 2024? Your customers, employees, and profits say ‘yes’
Related content: Does sustainability matter for businesses in 2024? Your customers, employees, and profits say ‘yes’
“Microsoft and other hyperscalers [large cloud service providers] are making commitments around 100% renewable energy in their data centres to address the sustainability challenges posed by AI,” said Nakazawa. “But as companies are sourcing and procuring those compute resources that are going to enable gen AI, they need to make sure that they’re doing the homework on whether their providers are really investing in more sustainable means.”
How businesses can optimize gen AI’s value without submarining their sustainability goals
In many ways, the gen AI die is already cast.
Adam Selipsky, CEO of AWS told Wired: “We do believe that generative AI will be transformative, will change the way that virtually every application in the world works, and will eventually transform the way that people work.”
This means that the responsibility of balancing the potentially enormous value of gen AI (or AI in general) with sustainability concerns will soon sit with executive teams across the world. If it hasn’t already.
So, how should they think about it?
Here are five key practices Nakazawa suggests should guide them:
1. First, understand the strategic value of gen AI
“Businesses first need to use generative AI in a way that’s as optimized as possible,” said Nakazawa. “But everybody’s still trying to figure out where and how gen AI can affect productivity, quality, customer experience, efficiencies and new business areas.”
By pinpointing the key areas of value for their particular organizations, businesses can make informed decisions about integrating AI effectively and precisely.
2. Only use gen AI intentionally
Nakazawa also advocates for “the intentional use of gen AI,” similar to how we turn off the lights in our homes when not in use to save energy. This minimizes unnecessary energy consumption (and associated costs).
3. Carefully evaluate suppliers’ sustainability commitments
Nakazawa suggests scrutinizing suppliers’ sustainability commitments, which should be transparent in their shareholder reports and annual disclosures. For instance, her company, NTT, has committed to investing $10 billion in data centre growth and achieving 100% renewable energy for its operations by 2030.
4. Monitor ongoing policy & regulatory trends
Because the AI policy landscape is still so young, Nakazawa recommends that companies monitor and engage with policies and regulations as they’re developed. That includes both the regulatory “stick” (penalties for non-compliance) and the “carrot” (incentives for meeting sustainability targets).
5. Collaborate with tech providers
Nakazawa recommends meeting with their technology providers to ensure their future updates and innovations support sustainable AI practices. This collaborative approach can help businesses stay at the forefront of (and influence the direction of) innovation while maintaining their sustainability goals.
The gen AI sustainability challenge will come down to company values
Ultimately though, Nakazawa believes how any given company navigates the AI sustainability challenges will come down to its mission, values, and employees.
“Increasingly, the younger generations are demanding that when they join a company, the company has to align with their values. And sustainability is a big topic with millennials and younger people. They’re going to care about the impact that AI has on sustainability and that’s going to make a difference inside companies.”
