Google's AI-driven emissions surge prompts sustainability debate

The rise of artificial intelligence (AI) is elevating carbon emissions to new heights, as demonstrated by recent news that Google's emissions have increased by nearly 50%, largely due to AI's energy demands. This revelation has sparked conversation among industry leaders about how to balance AI's burgeoning capabilities with the pressing need for sustainability.

Georges Olivier Reymond, CEO and founder of Pasqal has weighed in on the matter, offering insights into how quantum computing could potentially mitigate AI's environmental impact. He suggests pairing AI with quantum computing in a hybrid model, which he terms "Quantum AI," is a viable solution. "As AI datasets and usage grow, so does its energy consumption. Businesses need to find a way to harness its potential, just not at the cost of a sustainable future," Reymond stated.

Reymond emphasised that the unique advantages of quantum computers, which use qubits, allow for significant gains in computational efficiency without a corresponding rise in energy consumption. "Unlike traditional computers that use more energy with every added processor, quantum computers use qubits to achieve gains in the quality of results without requiring a similar amount of energy. This efficiency is particularly significant with newer types like cold atom machines, which are far more energy-friendly than their superconducting counterparts," he explained.

By integrating these technologies, businesses could boost the efficiency and speed of their AI algorithms and optimise energy usage, ensuring it is applied judiciously. "Combining AI with quantum computing can optimise algorithms and unlock new approaches to machine learning, all leading to more efficient AI. In doing so, energy usage is optimised to make sure it's only used when and where it's most needed," Reymond added.

Furthermore, he highlighted the quantum industry's focus on sustainability, noting that vendors are actively exploring solutions to accelerate the achievement of the UN's Sustainable Development Goals (SDGs) in critical areas such as health, energy, and climate protection.

Adding to the conversation, Dr Leslie Kanthan, CEO of TurinTech, warned that AI's environmental toll extends beyond initial training. "As businesses rush to harness AI's immense potential in products, processes, and services, from marketing to software to customer support, we can no longer ignore the technology's massive environmental toll. Significantly, AI's carbon footprint doesn't end at training; the emissions from ongoing inference to generate responses are also considerable," Kanthan emphasised.

Kanthan underscored the need for a comprehensive approach to ensure AI contributes positively to sustainability. "The AI industry needs wide-ranging, holistic strategies to make sure it's positively contributing to sustainability and not taking two steps back for every gain. Green energy, practical policies, and hardware transformations are essential," he stated. Additionally, Kanthan put a spotlight on code optimisation as an immediate, practical measure for reducing AI's carbon footprint. "Focusing on aspects like code optimisation can make a substantial difference for delivering immediate sustainable solutions for AI. And our planet can't afford to wait around for other options."

These comments highlight the growing consensus among industry leaders that sustainable AI is achievable and necessary. As organisations globally continue to integrate AI into their operations, the development and adoption of energy-efficient technologies like quantum computing and optimised code hold promise for mitigating the environmental impact while unlocking the full potential of AI.