The planet is heating up. Extreme weather events are on the rise. Climate change will exacerbate serious problems like drought, famine, biodiversity loss, and refugee crises; in some parts of the world it already has.
Could artificial intelligence be making things worse?
It's a question that Christina Shim, IBM's new chief sustainability officer, will have to grapple with in her new role. IBM offers its enterprise clients a variety of AI models, platforms, and tools, some of which aim to help them reduce their greenhouse gas emissions. But as AI has become the focal point of Silicon Valley's enthusiasm and investment, critics have scrutinized the technology's environmental downsides, especially the carbon emissions generated by AI server farms.
Google's latest environmental report, for instance, showed that the company's greenhouse gas emissions have grown 48 percent since 2019--in large part thanks to rising energy demands from data centers. Another recent study estimated that generating a single AI image is as energy-intensive as charging a smartphone.
"What does [it] mean when we have these big, significant technological breakthroughs that actually move us in the wrong direction?" asks Jennifer Layke, global director for energy at the World Resources Institute. "AI could be very useful for improving the long-term energy efficiency and integration of renewables into the energy system. But the question is really one of design and deployment strategy."
It's possible this technology will prove to be so helpful in the fight against climate change that its emissions will be worth it, Layke adds--but there's no such guarantee.
Inc. spoke with IBM's Shim about whether it's possible for firms to leverage AI without undercutting their sustainability goals, and why she remains optimistic about the technology's potential.
(This interview has been condensed and lightly edited.)
In the past few years, as AI--especially generative AI--has become a very mainstream thing for people and companies to talk about, there's been increasing criticism of its greenhouse gas emissions and other environmental impacts. Is there a way for companies to embrace this tech without exacerbating climate change?
It is possible, and I think it also has to be very intentional. I don't know that every company is thinking about it as intentionally as they should. There was a huge rush to think about "What does AI mean for my business? How do I make sure that I'm incorporating it?" without fully thinking through all the implications.
So a few key things here:
One is, how do you make sure that you're making the smart choice around your AI models? The biggest thing is, a bigger model is not necessarily a better model. Think about it from that vantage point: What is it that I need to achieve, and then what is the smallest model that I can use to get there? That helps with sustainability, but also costs and speed.
Two is, [at] IBM, we believe hybrid cloud [or a mix of on-site infrastructure with private and public cloud computing] helps in terms of locating your processing in an area that's close to clean energy. So for example, for us, 74 percent of our electricity consumed in our data centers came from renewable sources, and that's really important because even if you're having to use more of your data centers, at least you're using clean energy.
Three is around infrastructure. You really need to be investing in the right infrastructure. In 2023, for IBM, more than 58 percent of our energy conservation savings actually came from just upgrades in infrastructure, like our IT equipment in our data centers.
Are there mistakes or conceptual traps you see companies fall into that prevent them from fully embracing the environmental goals you're talking about?
Using a sledgehammer to hit a nail. Just going out and getting whatever generative AI model is out there--the one that the partner you happen to talk to happens to have--but not really making the smartest choice about that model and just going for whatever was available at the moment.
When people criticize AI's environmental impact, the folks building this technology often counter that AI will help us address climate change--that there are bigger-picture ways to use AI to find solutions or make things more efficient. Is that your perspective?
I think that's exactly right. I always ask, what is the lifecycle of the problem that we're trying to solve? So, yes, it might take more energy in the very beginning--but then, are we solving the problem faster?
If we're trying to solve for net zero for a particular company, can we do that in five years rather than 15 years? And then how much more are we saving over the course of the lifecycle of the problem?
IBM is heavily invested in the business side around AI for sustainability. We have a partnership with NASA building geospatial foundation models, using that to identify heat zones and putting policies in place to be able to adjust. We actually just did that with the Kenyan government and with the Emirati government. There are a lot of really powerful examples of how it can help with climate from a data management piece, from a geospatial piece; around physical assets, around the supply chain.
I take it you are generally an optimist about what AI means for the environment, and that the upsides outweigh the downsides. Why should ecoconscious people be optimistic about AI?
I don't think you can be in my job without being an optimist--AI aside. Even just the sustainability piece, you have to be.
Even if there are things that are pain points at this moment, there are also incredible efforts being put in place, including with some of my colleagues on the research side, to make sure that we are addressing some of that.
I never consider technology a savior for anything, but it helps to accelerate where we need to go, to get there faster--because we need to get there faster. And so I think I have to be an optimist, because if we're going to get to where we need to be, we need to make sure that we're leveraging technology as much as possible.
For further reading:
Download the IBM Gen AI Whitepaper for Financial Services
NOTE: This pose originally appeared on Inc.Magazine.com on August 12, 2024, and was written by Brian Contreras, Staff reporter, Inc. Magazine
#Sustainability, #GenerativeAI, #financial-services, #Featured-area-2, #Featured-area-2-home