show index hide index
The rise ofartificial intelligence is propelling this technology to the forefront with remarkable speed. However, this incessant quest for innovation involves significant challenges, particularly on the environmental level. The emergence and proliferation of AI-based solutions affects all sectors of activity, leading to staggering energy consumption which leaves an undeniable imprint on our ecosystem. What then is the current and long-term impact of this technological progression on our carbon footprint ?
The rapid advancement ofartificial intelligence (AI) is at the heart of our modernity, increasingly integrating all sectors of activity. However, its massive installation raises growing ecological concerns. This article explores how this promising technology, via its high energy consumption, affects the environment. It also highlights possible solutions to alleviate this impact while taking advantage of the benefits it offers.
An energy-intensive technological turning point
The development ofAI continues to impress with its ability to solve complex problems in various fields. However, the implementation of this technology requires a staggering amount of calculations, resulting in dramatic energy consumption. Training models such as GPT-3 consumed approximately 1,300 megawatt hours of electricity, an amount equivalent to the annual consumption of 120 homes.
Resource consumption: Greenhouse gases and water use
In addition to electricity, data centers dedicated to the operations of theAI also require large volumes of water for the cooling necessary for their proper operation. For training large models, water consumption reaches considerable levels, such as the 700,000 liters used by the GPT-3 infrastructure. In addition, the vast majority of this energy comes from fossil sources, contributing to carbon emissions. greenhouse gas.
Impact of daily use
The high consumption does not stop once the workout is over. Every interaction with AI systems like ChatGPT adds to the energy consumption. For example, a conversation with ChatGPT is equivalent to consuming 2.9 Wh, enough energy to power a 10 W LED bulb for 17 minutes. With the growing adoption of these technologies by the general public, this consumption is likely to increase significantly.
Solutions and prospects for improvement
Faced with these challenges, several solutions are being studied to make AI more ecological. Creating models that require less energy remains a focus. Furthermore, innovations like CartoLineBT from Enedis, which optimizes energy management, symbolize the growing interest in sustainable AI. These technologies make it possible to predict and effectively manage energy consumption, therefore reducing environmental impact in a contextual manner.
Compatibility with climate objectives
The rapid development of AI raises the question of its compatibility with the objectives of the Paris agreements, aimed at reducing global emissions. By 2026, total energy consumption linked to AI could reach 1050 terawatt hours, almost equivalent to Japan’s consumption. Forecasts indicate that CO2 emissions from AI could represent almost 0.9% of global emissions within a few years. It is therefore crucial to find a balance.
To read Giorgia Meloni : quand l’intelligence artificielle crée des images surprenantes en lingerie
Paradoxes and hopes
Despite its image as a dematerialized technology, AI has significant ecological impacts. Paradoxically, it may also be the key to solving some of these problems. By optimizing energy networks and predicting consumption needs, AI can contribute to more efficient management and reduce energy waste in other sectors.
