Today, Salesforce have announced the release of the AI Energy Score, working alongside Carnegie Mellon University, Cohere, and Hugging Face to release a first-of-its-kind benchmarking tool.
As concerns grow larger around sustainable AI in the United States, this announcement highlights Salesforce’s position as a true advocate for energy-efficient AI models.
What Is the AI Energy Score and Why Is It Important?
Significant amounts of energy must be used to power artificial intelligence models. By 2027, AI could use between 85 to 134 terawatt-hours of electricity annually – roughly equivalent to the power consumption of a small country like the Netherlands.
The high energy footprint of AI raises concerns about its impact on the environment, and researchers are pushing large companies, as well as the government, to prioritize sustainable AI.
This new AI Energy Score helps businesses measure and optimize how much energy their AI is consuming, giving a clear picture of cost, efficiency, and sustainability.
Salesforce has long been vocal about its commitment to environmentally friendly AI practices, and the AI Energy Score further encourages transparency in efficiency standards, establishing a clear and trusted benchmark for AI model sustainability.
“Reducing AI energy consumption lowers operational costs, optimizes infrastructure, and enhances long-term sustainability and profitability,” said Suzanne DiBianca, EVP and Chief Impact Officer of Salesforce. “We are proud to work with industry leaders to build a more transparent AI ecosystem.”
The AI Energy Score features:
- Standardized Energy Ratings: A framework for measuring and comparing AI model energy efficiency.
- Benchmarking Portal: AI developers can submit open or proprietary (closed) models to be tested by others.
- Public Leaderboard: A leaderboard that scores for ten common AI tasks – from text generation to summarization.
- Energy Use Label: A one-to-five star rating system to help developers identify the best-performing models.
Once rated, developers can create a standardized label to display their energy score on their AI model.
This is a real game-changer for how other businesses can use AI, encouraging people to make smarter decisions about how they’re using AI without overspending on computing power.
Companies are now under more pressure to meet sustainability targets, and AI’s growing energy consumption is a hidden challenge. This scoring system helps companies strategize how they’re going to combat this.
Ariane Thomas, Global Tech Director of Sustainability of L’OREAL Group, said: “Transparency like that offered by the AI Energy Score is crucial. This initiative sets an important milestone, and we hope more proprietary model providers join us in adopting such practices to drive accountability and sustainability across industries and supply chains.”
“By openly sharing energy consumption data, our companies can collectively implement ecodesign practices and minimize the environmental footprint of AI technologies.”
Industry leaders will get a first glimpse at the Energy Score at the AI Action Summit in France this week.
Final Thoughts
AI services and systems are growing at an unprecedented rate. While this is equally exciting as it is worrying, it’s important that environmental factors don’t get swept under the rug, as it’s something that could easily lose control as countries accelerate their AI efforts.
Salesforce’s commitment to efficient AI is something the ecosystem can take pride in, as its initiatives have the potential to set a lasting precedent for how other large-scale companies adopt and optimize AI.
