Salesforce And Partners Launch AI Energy Score For Sustainability

Salesforce, in collaboration with AI community platform Hugging Face, AI developer Cohere, and Carnegie Mellon University, has unveiled the AI Energy Score, a new benchmarking tool designed to help AI developers and users evaluate and compare the energy consumption of AI models. The initiative comes amid growing concerns about the environmental impact of AI, as its rapid adoption threatens corporate carbon reduction efforts. According to the International Energy Agency, artificial intelligence may well double the data centre power usage by 2026, which represents a gigantic challenge for tech industry giants like Microsoft, Google, and Amazon, striving to reach decarbonization goals.

A recent Salesforce survey revealed that sustainability professionals are grappling with the environmental implications of AI, even as they recognize its potential to benefit supply chain management, ESG reporting, and regulatory compliance. Addressing this dilemma, the AI Energy Score aims to provide transparency around the environmental footprint of AI models, similar to the ENERGY STAR label for appliances and electronics.

The tool offers a standardized rating framework to measure and compare AI model energy efficiency. It features a public leaderboard showcasing scores for 166 commonly used AI models and a portal for developers to submit their models for evaluation. Additionally, AI models will receive energy use labels rated from 1 to 5 stars to guide developers toward more sustainable options.

Salesforce announced that it will be the first model developer to disclose the energy efficiency of its AI models using the new framework. Suzanne DiBianca, EVP and Chief Impact Officer at Salesforce, highlighted the initiative’s importance, stating that reducing AI energy consumption not only lowers operational costs and optimizes infrastructure but also enhances long-term sustainability and profitability.

Dr. Sasha Luccioni, AI & Climate Lead at Hugging Face, described the AI Energy Score as a significant milestone for sustainable AI, emphasizing the importance of a transparent rating system in addressing the environmental impact of AI. She expressed optimism about the broader adoption of the tool, which aims to drive market preference for energy-efficient models and incentivize sustainable AI development.