Generative AI famously consumes large amounts of energy due to the extensive computational power required for training models and processing tasks, and this is raising concerns about its contribution to environmental issues such as increased carbon emissions and resource depletion.
Sasha Luccioni, a prominent Canadian computer scientist and AI researcher known for her work on the environmental impact of artificial intelligence, recently discussed the topic at the ALL IN artificial intelligence conference in Montreal.
“I find it particularly disappointing that generative AI is used to search the Internet,” Luccioni told AFP on the sidelines of the conference, lamenting the energy costs of using AI for tasks that traditional search engines could perform.
Thoughtful, efficient use of AI
Unlike basic search engines that retrieve existing information, AI models generate new content, requiring significant computing power to train on billions of data points and respond to user requests. Google’s AI Overview feature, for example, offers AI-generated snapshots that summarize key points from multiple sources in response to complex queries.
“We are accelerating the climate crisis,” Luccioni warned, calling for more transparency from tech companies and urging governments to legislate more effectively once such transparency is achieved.
Recognized by Time in 2024 as one of the 100 most influential figures in AI, Luccioni helped create a tool called “CodeCarbon” in 2020, which enables developers to measure the carbon footprint of their code. This tool has been downloaded over a million times. She is now working on a certification system to label AI models based on their energy efficiency. Comparing the system to the US Environmental Protection Agency’s energy consumption ratings for appliances, she explained: “For a specific task, we can measure energy efficiency and say that this model has an A+, and that model has a D.”
While she might be perceived as anti-AI by some, Luccioni is keen to stress that her goal is to promote “energy sobriety” by encouraging users to make thoughtful, efficient use of AI technologies.
AI is already better at analyzing financial reports than humansAI hype is at an all time high, but only a few of us actually use it daily Why controlled experimentation is a key step in an AI roll-out