Critics have hitherto cause concern About the energy consumption of generative AI models that power chatbots such as ChatGPT and Bard. But we’ve been here before, according to to A new report by the Information Technology Innovation Foundation (ITIF), a nonprofit think tank. Near the peak of the dot-com boom in the 1990s, Forbes article “Somewhere in America, a lump of coal is burned every time a book is ordered online,” he lamented.
The authors of this article predicted that within 10 years, half of the electricity grid will be powering the Internet economy. After all, they weren’t even close. The International Energy Agency (IEA) estimates that today’s data centers and data transmission networks “each account for approximately 1 to 1.5 percent of global electricity use.”
Generative AI models require: vast computing power To create new content. However, as with past technologies, many of the early claims about AI’s energy consumption have proven to be “exaggerated and misleading,” the ITIF report said.
So what’s different this time? According to the authors, AI can help mitigate climate change.
The challenge of estimating AI energy consumption
Accurately estimating the energy usage and carbon footprint of an AI system over its lifetime is difficult. Because these calculations depend on many factors, including the details of the chip, cooling system, data center design, and energy source.
Most large-scale AI models require more energy than small-scale AI models. For example, Google’s electricity consumption has increased as its business has grown, especially from its data centers. The tech giant’s data centers used about 3 terawatt-hours more electricity in 2022 than the previous year. However, while Google’s overall energy usage has increased, the percentage devoted to machine learning has remained constant at between 10% and 15%.
and the energy required for AI model inference, the process of feeding data to a trained model. Make predictions and solve tasksAccording to the ITIF report, there is an overall decline with each new chip release.
AI addresses climate change
This report details how AI can help reduce energy consumption in several industries, including transportation, agriculture, and energy. The technology can interpret complex climate data from sensors and satellites, such as sea level changes and rainfall, to create better forecasts and address the risks of climate change. Farmers have also been using AI for precision farming for years to reduce fertilizer and water usage.
Businesses and governments are leveraging AI to run buildings, roads, and waterways more efficiently. In California, for example, the government has implemented it to detect and quickly respond to wildfires, reducing carbon emissions from fires. Meanwhile, logistics companies are leveraging AI to optimize delivery routes and reduce fuel consumption.
The ITIF report concludes that all activities that use energy have an impact on the environment, and AI is no exception. However, the researchers found that there are no “market-specific failures” related to the technology’s power consumption that would have a greater impact than in alternative applications. For example, a kilowatt-hour used for AI is no different than a kilowatt-hour used to watch TV or microwave popcorn.
