Google recently announced their new Search Generative Experience (SGE) at their I/O conference, planning to integrate generative AI into Google Search and deliver AI-generated answers to over a billion users by 2024. While this could enhance the user experience by providing richer context to query results, concerns have been raised over the increased computational power and environmental impact of such functionality.
AI-based searches are predicted to require more energy, with researcher Sasha Luccioni estimating that using AI to generate search information could need 30 times the energy of traditional searches. Given that billions of Google searches are carried out daily, the environmental impact could be substantial, especially with the energy-intensive process of training AI models.
Google’s SGE also poses significant infrastructure challenges. Data centers could double their energy consumption by 2026 due to AI, necessitating massive power inputs which have been likened to multiple nuclear power stations’ output. As such, companies are keen to highlight the efficiency of AI hardware, and the rising capability of renewable energy. However, transitioning to renewable energy in data centers is a complex and slow process, with the intermittent nature of wind and solar energy falling short of the constant power required by these centers.
The spike in energy consumption for AI operation also extends to water. Data center water usage related to the generative AI industry could surge. Microsoft, for instance, recorded a rise in water usage of up to 50% in some areas, with AI being identified as the major driver of this consumption. Studies also suggest that generative AI’s water footprint could be massive, potentially equivalent to the daily consumption of a city of over half a million people.
Although promising improvements in search efficiency and user experience, Google’s SGE also throws up serious questions around sustainability. With a projected increase in energy and water consumption, the environmental footprint of AI could escalate significantly, posing issues for the industry’s sustainability in the future. As the adoption and integration of AI continue to increase, companies will need to continue finding innovative and scalable solutions to manage the environmental impact effectively.
