2023 has been defined by artificial intelligence (AI) – McKinsey calls it AI’s breakout year, with their survey showing one-third of all respondents are using generative AI (gen AI) regularly in at least one business function. This is unsurprising, given the massive efficiency and productivity gains that come with gen AI. Asia, in particular, has emerged frontrunner in the global race to adopt gen AI. A separate survey by EY noted that nearly all business leaders surveyed in the region have already invested, or are planning to, significantly in the technology.
The optimisation and efficiency benefits unlocked by AI have become so prominent that even discussions at this year’s COP28, a critical climate change-focused conference, cast a spotlight on AI’s role in alleviating our climate issues. The discussions around AI’s role in combating climate change, and the number of high-profile partnerships on AI-centric solutions resulting from the summit bear strong testament toward AI as one of the “heroes” in the race to build a more sustainable world.
However, as with all powerful tools, there's a darker side to consider: AI's insatiable hunger for computing power. As AI models become more complex, their insatiable thirst for data and processing capacity translates to a growing carbon footprint. A recent study paints a stark picture: AI servers could consume up to 0.5% of the world's current electricity use by 2027 – about the same as several small countries combined. For Asia – already the world’s largest carbon emitter – its growing appetite for AI could spell trouble if adoption is not balanced with one eye on its impact on the environment.
Could AI then be the solution to climate change that we have been looking for, or is it just one more source of carbon output that will simply accelerate our ongoing climate disaster?
AI as a catalyst for climate change – the downside
While AI is expected to be one of the “heroes” in the fight against climate change, some businesses have shared their hesitation toward the widespread use of the technology. Chief among their concerns is the appetite AI has for electricity, given the immense computing power required to run these systems.
See also: 80% of AI projects are projected to fail. Here's how it doesn't have to be this way
As AI models become more complex in terms of programming and design to account for new datasets and new ways to process those datasets, the amount of computing power the average model needs exponentially increases. A peer-reviewed analysis published in October finds that AI servers could use between 85 to 134 terawatt hours (TWh) annually by 2027, about 0.5% of the world's current electricity use.
Furthermore, the Apac market size for AI is only set to expand exponentially, growing at a CAGR of over 39% to reach US$254.81 billion by 2028. All this growth necessitates a relook at conversations that speak only of the upsides of AI, without consideration for its very real and present deficiencies.
This, then, is the green double-edged sword we hold – a potent tool for good, with a hefty potential for environmental harm. So, how do we wield it responsibly? How can we harness the power of AI for good and achieve our climate goals?
See also: Responsible AI starts with transparency
How AI can alleviate climate change and carbon emissions – the upside
The allure is undeniable. AI's ability to optimise processes, analyse mountains of data, and drive efficiency holds immense promise for a greener future. Imagine smart grids anticipating energy needs and distributing renewable power with laser precision. Envision self-optimising factories minimising waste and emissions, driven by AI algorithms crunching numbers in real-time. Think of AI-powered logistics streamlining transportation networks, slashing travel distances and carbon footprints. These are just the tip of the iceberg, mere glimpses of the sustainable paradise AI could pave the way for.
A recent survey by Gartner revealed that a wide range of CEOs believe AI is the main technology that will significantly impact their industry over the next three years. It also highlighted the growing awareness amongst today’s business leaders about AI’s expanding carbon footprint and the need to have balanced perspectives around AI — not just pursue growth and adoption blindly. It is no wonder that 78% of surveyed CEOs said the benefits of AI outweigh the risks.
With a potentially strong positive upside, and also very real downsides to increased adoption of AI across industries, is there a balance that can be struck, and if so, what could it look like?
Achieving a positive balance between scale and emissions
While there may not be an easy solution to striking a perfect balance between the scale of AI adoption and its emissions, the positives of AI outweigh the negatives when it comes to achieving our sustainable goals and AI development should therefore not be stymied.
In terms of power consumption, it is important to note that the world is pushing for a greater variety of renewable energy sources and that more than a third of the world’s power sources are expected to become renewables in two years. The world's shift towards renewables is not just a climate imperative, but also a prerequisite for powering our AI revolution sustainably. Investing in solar, wind, and other clean energy sources will decouple AI's growth from fossil fuel dependence, paving the way for a future where technology and nature live in harmony.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
Green AI is lean and efficient AI
AI’s duality as both a powerful ally and a potential polluter is clear. Blindly chasing its benefits while ignoring its downsides is a recipe for disaster. Transparency and responsible development are key. The recent signing of the world’s first pact for safe AI is an encouraging development in this direction, as are individual governments’ efforts to exercise AI governance alongside efforts to stimulate innovation. Singapore, for instance, just launched its revised National AI strategy in effort to scale adoption while mitigating risks. Indonesia, too, issued its own local set of guidelines in December.
Even still, regulatory governance is just one part of the puzzle. We must simultaneously be meticulous in designing AI models with sustainability in mind, prioritising efficiency over raw power whenever possible. Think lean algorithms, energy-efficient hardware, and a laser focus on maximising impact with minimal resources. After all, as intelligent as AI is, we are still ultimately responsible for its use, and it is up to business leaders to carefully calibrate their rates of AI adoption with the technology’s overall energy consumption.
A PwC report forecasts that the application of AI in various sectors could reduce worldwide greenhouse gas (GHG) emissions by four percent in 2030, equivalent to the 2030 annual emissions of Australia, Canada, and Japan combined. However, to achieve this sort of outcome, we need to encourage collaboration between the public and private sectors. And, within the technology industry we need to foster innovation not just in AI hardware, but software as well, from developing energy-efficient processors, to optimising algorithms for reduced resource consumption and more.
The path forward: Embracing shared responsibility
The path forward is not without challenges. Balancing AI's immense potential with its environmental footprint demands careful navigation, bold innovation, and collective responsibility. But the rewards are too vast to ignore.
By embracing the green double-edged sword of AI responsibly, we can unlock a future where technology serves as a bridge towards a more incredible, sustainable world. With Asia having as much to gain from AI as it has to lose from accelerated climate change, this will be more urgent than ever.
Samson Hu is the co-chief executive officer of ASUS
