The tech industry needs to do more to address growing concerns about the sustainability of AI, according to NetApp CEO George Kurian.
Talking with IT Professional At the company's NetApp Insight conference in Las Vegas, Kurian shared his thoughts on the environmental impact of how companies are deploying technology.
Asked how leaders are balancing sustainability goals with increasing pressure to adopt energy-intensive systems, Kurian insisted that companies need to improve the way they deploy AI technologies and ensure they do so intelligently to minimize unnecessary energy use.
“I think the world needs to do a better job of thinking about how we can use AI in a way that delivers the best outcomes,” he said. IT Professional.
The recent explosion of generative AI has created an environment where companies are forced to deploy the technology at a rapid pace to ensure they don’t fall behind their competitors. But this comes with the risk of deploying the technology in a haphazard manner, Kurian warned, creating long-term challenges.
To counter this, NetApp’s CEO said businesses should spend more time formulating sound data strategies to avoid future problems. When they get it right, Kurian added, companies will be able to drive more sustainable operations and deliver noticeable cost benefits.
“When it comes to AI, it’s important to use data intelligently because if you’re trying to search through large amounts of data for a small amount of benefit, you’re wasting a huge amount of expense and energy,” he explained.
“Many of the data management tools we are developing allow you to use the data very precisely and, as the data changes, use the model only for the changes in the data. This is beneficial for costs, time to value and sustainability.”
AI creates huge amounts of hidden data that continues to consume energy
Talking with IT Professional, Nicola Acutt, NetApp’s chief sustainability officer, said the biggest obstacle to AI sustainability was the data challenge. Acutt believes the biggest driver of energy intensity for many enterprise AI applications is “dark data,” also known as “shadow data.”
Dark data is information that is collected, processed, and stored, often before the company ever uses it again.
This type of data is often backed up or stored in a variety of forms, such as unstructured files, databases, or even cloud storage. Crucially, shadow data is rarely properly managed or maintained, and IT teams often have no idea whether it contains potentially sensitive information.
Many companies use this type of data to train AI systems on a one-time scenario, Acutt explained, and then discard it despite the fact that it still consumes vital resources.
“What’s amazing about this is that almost two-thirds of the data that’s created for AI today is created but then never used again… but it still consumes energy and ultimately that translates into a large carbon footprint from the data that we don’t use again,” Acutt said.
He warned that companies will have to confront this reality and take responsibility for the data they collect, and advised IT leaders to re-evaluate governance approaches.
“There will be a need to recognize the scope of dark data and there will be a huge need for intelligence around it,” he explained.
AI means organizations will fail to meet their sustainability goals
Jon Brown, senior analyst at Enterprise Strategy Group (ESG), said IT Professional Their own research indicates that some industries are implementing AI better than others, and that some felt it would boost sustainability.
“We saw a lot of different opinions about whether people would be able to meet their commitments and whether AI would have a positive, mixed or negative effect on sustainability,” he said.
“What we found was that for a couple of verticals, they were overwhelmingly positive about the impact AI was going to have on their sustainability commitments, like retail, very optimistic; technology companies like NetApp, very optimistic that AI would be a net positive for [sustainability].”
But some industries were less confident, Brown noted, suggesting that implementing generative AI in their organization would jeopardize their ESG goals.
“When we started looking at manufacturing and some of the other verticals out there, they said it was going to be a mixed bag, 'we're going to have to balance this and recognize that it's going to be a challenge to meet these commitments.'
Many large organizations set ambitious sustainability goals, such as net zero by 2030, in the years leading up to 2020, but the frenzy around generative AI sparked by OpenAI’s launch of ChatGPT in November 2022 put these goals on shaky ground.
“I see some of those sustainability directors retiring in 2029,” Brown joked, adding that he expects several of these companies to quietly revise their targets to reflect more realistic reductions.
The review of ambitious ESG targets has already begun at some major providers.
Earlier this year, both Microsoft and Google warned that their increased focus on developing generative AI had resulted in a sharp increase in data center energy consumption.
Google's total greenhouse gas emissions According to official figures released by the tech giant, CO2 equivalent emissions in 2023 amounted to 14.3 million tonnes, up 13% year-on-year and 48% higher than in 2019, when its sustainability targets were first set.
Similarly, Microsoft's 2024 plan Sustainability Report I found your Carbon emissions increased by 29% as a result of its drive to expand data center capacity.
The problem has reached such a point that reports from July showed Microsoft and Google's combined electricity consumption exceeded that of about 100 countrieshighlighting the sharp increase in energy demands that generative AI places on infrastructure.