New AI Energy Score Tool Ranks AI Models by Sustainability

As artificial intelligence continues to drive up energy consumption worldwide, a new online tool is giving developers and companies a way to evaluate the environmental impact of their AI models. Salesforce introduced AI Energy Score at the AI Action Summit in Paris on Monday, marking a significant step toward increasing transparency in AI's energy use.

The tool allows users to compare the energy efficiency of various AI models, helping companies make more sustainable choices. By offering standardized energy ratings and a public leaderboard highlighting top-performing models for common tasks like image creation, text generation, and summarization, AI Energy Score aims to guide users toward more eco-friendly AI solutions.

Promoting Transparency in AI Energy Use

Salesforce developed AI Energy Score in partnership with Canadian AI company Cohere, Carnegie Mellon University, and the tech platform Hugging Face. This collaboration brings together expertise from industry leaders, academic researchers, and AI transparency advocates to create a comprehensive tool for evaluating AI’s environmental impact.

The tool features a recognizable star rating system, similar to the Energy Star ratings for home appliances, providing an easy-to-understand assessment of an AI model’s efficiency.

Boris Gamazaychikov, Senior Manager of Emissions Reduction at Salesforce, emphasized that sustainability officers are often left in the dark about the environmental impact of the AI models their companies use.

"The sustainability officers are kind of in a bind right now, there isn't a lot of transparency yet with these external models," Gamazaychikov told Newsweek. "They're seeing their companies use kind of these black box models and not really know what kind of impact that's going to have on their carbon footprint."

AI’s Growing Energy Demand and Environmental Impact

AI requires significant computing power, both during the training phase and for daily use (known as inferences). This demand places a growing strain on global energy resources. A December report from the U.S. Department of Energy predicted that data center energy use could triple by 2028, accounting for up to 12% of the nation’s total electricity consumption.

This surge in energy use has already caused tech giants like Microsoft and Google to miss their climate targets, while also threatening to exacerbate greenhouse gas emissions and other forms of pollution. A separate study published in December warned that the rise of AI could increase health risks from air pollution unless data centers switch to cleaner energy sources.

A Push for Industry-Wide Change

Sasha Luccioni, AI and climate lead at Hugging Face, has long advocated for greater transparency in AI’s environmental footprint. She likened the AI Energy Score to the Energy Star rating system, offering consumers a clear, standardized way to compare efficiency.

"Part of that problem is that AI is such an umbrella term that people don't really know what we're talking about," Luccioni said, emphasizing that standardized measurements will give companies and consumers a better understanding of the energy implications tied to AI use.

Luccioni hopes that the tool will encourage tech companies to be more transparent about their energy consumption. "I think it's going to help companies open up a little," she added.

Gamazaychikov echoed this sentiment, expressing hope that the AI Energy Score will nudge tech companies toward more sustainable practices.

"That's kind of the big swing, the big hope with all this," he said. "We are trying to tackle it from that perspective, trying to drive transparency and thus more efficiency in the space."

What This Means

AI Energy Score represents a crucial step in addressing the environmental consequences of artificial intelligence. As AI becomes more deeply embedded in industries worldwide, tools like this can guide developers and companies toward more sustainable choices by providing transparent, easy-to-understand metrics on energy use.

However, the broader challenge remains: ensuring that the rapid growth of AI doesn’t come at the expense of the planet. The energy demands of AI models are set to increase dramatically, and without widespread adoption of clean energy solutions and transparent reporting, the environmental costs could outweigh the benefits.

This tool, while a positive step, is only part of the solution. The tech industry will need to prioritize sustainable development and embrace innovations that reduce energy consumption if it hopes to meet global climate goals.

Editor’s Note: This article was created by Alicia Shapiro, CMO of AiNews.com, with writing, image, and idea-generation support from ChatGPT, an AI assistant. However, the final perspective and editorial choices are solely Alicia Shapiro’s. Special thanks to ChatGPT for assistance with research and editorial support in crafting this article.