Researcher launches Sustainable AI Group to push AI emissions data

In short: A leading AI sustainability researcher is starting a new group to help companies measure and reduce the environmental impact of AI, and to push big AI providers to share clearer emissions data.

What happened

Sasha Luccioni, an AI sustainability researcher, is launching Sustainable AI Group with Boris Gamazaychikov, a former sustainability leader at Salesforce. The goal is to help companies understand and improve the environmental footprint of the AI they use.

Luccioni is known for pushing for more transparency about AI energy use and emissions. At Hugging Face, she helped create a public leaderboard that compares how energy-efficient different open-source AI models are. Open-source means the model’s “recipe” is shared publicly, so others can study it and build on it.

In an interview with Wired, Luccioni said many companies are facing pressure from employees and executives to “quantify” AI’s impact. She argues businesses need basics like where their AI systems run, which data centers are used, and what energy powers those data centers. A data center is a warehouse filled with computers that run online services.

Luccioni also said AI companies should show users simple energy information, like a “meter” on a chatbot that tells you how much energy a conversation used, and how much pollution it caused (like a receipt after you check out). She also wants more research on the energy needs of different AI tasks, such as speech-to-text and photo-to-video.

Why it matters

AI use is growing fast, and it often depends on large data centers that can use a lot of electricity, sometimes from fossil fuels. Without clear numbers, it is hard for companies, governments, and everyday users to make informed choices, like picking smaller models when they are good enough or paying more for cleaner power.

Source: Wired