The year 2023 has been marked as the breakout year for artificial intelligence (AI), according to McKinsey, with a third of respondents regularly employing generative AI (gen AI) in business functions. Particularly noteworthy is Asia’s lead in the global adoption of gen AI, as indicated by a survey from EY, revealing that nearly all business leaders in the region have invested or plan to invest significantly in this transformative technology.
The efficiency and productivity gains associated with gen AI have propelled it into the spotlight, even influencing discussions at COP28, a critical climate change-focused conference. The attention on AI’s role in addressing climate issues has led to high-profile partnerships and initiatives emerging from the summit, underscoring AI’s potential as a pivotal player in the pursuit of a more sustainable world.
However, amidst the optimism, a darker aspect of AI’s influence comes to the fore—the substantial demand for computing power. As AI models become more intricate, their hunger for data and processing capacity raises concerns about a burgeoning carbon footprint. A recent study forecasts that AI servers could consume up to 0.5% of the world’s current electricity usage by 2027, a figure comparable to that of several small countries combined. For Asia, already the largest carbon emitter globally, the increasing appetite for AI poses a potential environmental threat if not balanced with considerations for its impact.
The question arises: Can AI be the solution to climate change, or does it risk exacerbating our ongoing environmental challenges?
AI as a Climate Change Catalyst: The Downside
While AI is positioned as a hero in the fight against climate change, some businesses express reservations about its widespread use. The primary concern revolves around AI’s substantial electricity consumption, driven by the escalating computing power required for increasingly complex models. A peer-reviewed analysis suggests that AI servers could consume between 85 to 134 terawatt hours annually by 2027, amounting to 0.5% of the world’s current electricity usage.
The expanding market size for AI in the Asia-Pacific region, projected to reach US$254.81 billion by 2028, demands a reevaluation of discussions solely focusing on AI’s positive aspects without acknowledging its present deficiencies.
This sets the stage for a delicate balance between AI’s potential for good and its significant environmental impact. How can we responsibly wield this green double-edged sword and leverage AI for positive climate outcomes?
Harnessing AI for Climate Solutions: The Upside
The allure of AI’s potential to optimize processes, analyze vast datasets, and drive efficiency presents promising avenues for a greener future. Imagine smart grids anticipating energy needs, self-optimizing factories minimizing waste, and AI-powered logistics streamlining transportation networks—these scenarios offer glimpses of the sustainable paradise AI could help create.
A recent Gartner survey revealed that CEOs see AI as the main technology influencing their industries over the next three years. The growing awareness among business leaders about AI’s expanding carbon footprint is evident, with 78% of surveyed CEOs acknowledging that the benefits of AI outweigh the risks.
While the downsides of increased AI adoption are evident, there is no denying the positive potential AI holds for achieving sustainability goals. Striking a balance between scale and emissions becomes crucial in this context.
Achieving a Positive Balance: Green AI and Responsible Development
Though finding a perfect equilibrium between the scale of AI adoption and its emissions poses challenges, the positive aspects of AI outweigh the negatives when it comes to sustainable goals. Power consumption considerations highlight the global shift toward renewable energy sources. Over a third of the world’s power sources are expected to be renewables in two years, providing a climate-friendly foundation for the AI revolution. Investments in solar, wind, and other clean energy sources will decouple AI’s growth from fossil fuel dependence, fostering a future where technology and nature coexist harmoniously.
The concept of “Green AI” emphasizes lean and efficient AI development. Transparent and responsible practices, such as the recent signing of the world’s first pact for safe AI and government initiatives like Singapore’s revised National AI strategy, play crucial roles in guiding AI development.
While regulatory governance is essential, meticulous design of AI models with sustainability in mind is equally vital. Prioritizing efficiency over raw power, using lean algorithms, energy-efficient hardware, and maximizing impact with minimal resources are key strategies. A PwC report suggests that AI applications could reduce worldwide greenhouse gas emissions by 4% in 2030, equivalent to the annual emissions of Australia, Canada, and Japan combined.
The path forward demands shared responsibility, bold innovation, and careful navigation to balance AI’s immense potential with its environmental impact. Embracing the green double-edged sword of AI responsibly can pave the way for a future where technology acts as a bridge to an incredible, sustainable world—a necessity for Asia, which has as much to gain from AI as it stands to lose from accelerated climate change.