by Artificial intelligence (AI) has the potential to revolutionize various industries, from coding to autonomous driving. However, a recent commentary published in the journal Joule suggests that widespread adoption of AI could come with a significant energy footprint, potentially surpassing the power demands of certain countries.
The study, conducted by Alex de Vries, founder of Digiconomist, examines the growing demand for AI services and predicts a substantial increase in energy consumption related to AI in the coming years. One factor contributing to this rise is the rapid growth of generative AI, which can produce text, images, and other data. The training process for these AI models requires a significant amount of energy-intensive data input. For instance, Hugging Face, an AI-developing company, reported that its multilingual text-generating AI tool consumed about 433 megawatt-hours (MWH) during training, equivalent to the power needed to sustain 40 average American homes for a year.
Moreover, the energy consumption does not end with training. De Vries highlights that each time a generative AI tool generates text or images based on prompts, it utilizes a significant amount of computing power and, consequently, energy. For example, ChatGPT, developed by OpenAI, could consume approximately 564 MWh of electricity daily.
Although efforts are being made to improve the efficiency of AI hardware and software, de Vries argues that increased efficiency often leads to an increase in demand, known as Jevons’ Paradox. This suggests that technological advancements will ultimately result in a net increase in resource use, potentially offsetting any energy-saving measures.
Companies like Google are already integrating generative AI into their services, such as email and search engines. De Vries estimates that if every Google search were powered by AI, it would require approximately 29.2 TWh of power annually, equivalent to Ireland’s annual electricity consumption. While this extreme scenario is unlikely to happen in the near future due to cost and supply chain limitations, the production of AI servers is projected to increase significantly. Based on this projection, global AI-related electricity consumption could rise by 85 to 134 TWh annually by 2027.
The potential growth in AI energy consumption raises concerns about the thoughtful use of this technology. De Vries emphasizes the need to consider whether AI is truly necessary in various applications, given its energy-intensive nature. As improvements in AI efficiency continue, there may be a further increase in AI-related electricity consumption, particularly if computer processing chips are repurposed for AI use.
In conclusion, while AI offers numerous advantages, its energy requirements cannot be ignored. It is crucial to carefully consider the applications of AI to avoid unnecessary energy consumption.
*Note:
1. Source: Coherent Market Insights, Public sources, Desk research
2. We have leveraged AI tools to mine information and compile it
Money Singh is a seasoned content writer with over four years of experience in the market research sector. Her expertise spans various industries, including food and beverages, biotechnology, chemicals and materials, defense and aerospace, consumer goods, etc.
