If you’re concerned about the environment, consider the AI environmental cost before using generative artificial intelligence. Researcher Sasha Luccioni highlights that generative AI consumes 30 times more energy than traditional search engines, raising critical awareness about the environmental impact of this emerging technology.
Recognized as one of Time magazine’s 100 most influential people in AI in 2024, Sasha Luccioni, a Canadian computer scientist of Russian descent, has been dedicated to quantifying the emissions associated with AI programs like ChatGPT and Midjourney for several years.
Sasha Luccioni expressed disappointment that generative AI is increasingly used for online searches. Speaking at the ALL IN artificial intelligence conference in Montreal, she noted that the language models powering these programs demand substantial computing resources, trained on billions of data points that require powerful servers.
In 2022, the combined electricity consumption of the AI and cryptocurrency sectors reached nearly 460 terawatt hours, accounting for 2% of total global electricity production. Unlike traditional search engines that simply retrieve information, AI programs generate new content, making them significantly more energy-intensive.
As a leading climate impact researcher, Sasha Luccioni co-developed CodeCarbon in 2020, a tool designed to help developers quantify the carbon footprint of their code. With over a million downloads, this tool exemplifies the growing need for energy efficiency in AI.
Currently, Sasha Luccioni heads the climate strategy at Hugging Face, a platform for sharing open-access AI models. She is working on creating a certification system for algorithms to measure their energy consumption, similar to the US Environmental Protection Agency’s program that rates electronic devices based on energy efficiency. “This allows us to assess models on a scale, giving one an A+ and another a D,” she explains, acknowledging that water use and rare materials are not considered.
To enhance transparency, Sasha Luccioni is testing her tool on open-source generative AI models but seeks collaboration with commercial entities like Google and OpenAI, which have been reluctant. Despite commitments from Microsoft and Google to achieve carbon neutrality by the end of the decade, their greenhouse gas emissions surged in 2023—up 48% for Google compared to 2019 and 29% for Microsoft since 2020.
“We are accelerating the climate crisis,” Sasha Luccioni warns, urging tech companies to provide greater transparency. She believes government intervention is crucial, as they currently lack insight into data sets and algorithm training processes. “Once we have transparency, we can start legislating.”
Also Read: First AI Exhibition in South Punjab to Showcase Tech Innovations
Promoting “energy sobriety” is essential, Sasha Luccioni emphasizes. It’s not about opposing AI but choosing the right tools and using them wisely. Her research indicates that generating a high-definition image with AI can consume as much energy as fully recharging a smartphone battery.
Also Read: Meta Trains AI with Facebook & Instagram User Data
As businesses increasingly seek to integrate AI into everyday life through conversational bots, connected devices, and online searches, it is imperative to consider the AI environmental cost and the choices we make in leveraging this powerful technology.
