With AI use skyrocketing, Google has decided to break its silence on one of the technology’s most controversial factors: environmental impact. In a blog post titled “Measuring the impact of AI inference,” Google breaks down its methodology for measuring the energy, emissions, and water impact of its Gemini prompts.
It’s important to note that Google has a lot at stake here, with the company going all-in on its Gemini AI products across all of its services. The company has also released a technical paper that goes much more in-depth on the specifics, but its conclusion is as follows:
“Using this methodology, we estimate the median Gemini Apps text prompt uses 0.24 watt-hours (Wh) of energy, emits 0.03 grams of carbon dioxide equivalent (gCO2e), and consumes 0.26 milliliters (or about five drops) of water — figures that are substantially lower than many public estimates,” says Google. “The per-prompt energy impact is equivalent to watching TV for less than nine seconds.”
Not all experts are convinced, with many pointing out that Google does not consider indirect water usage in its calculations.
“They’re just hiding the critical information,” Shaolei Ren, an associate professor of electrical and computer engineering at the University of California, Riverside, tells The Verge. “This really spreads the wrong message to the world.”
For Google’s part, the company says that it isn’t resting on its laurels, and is “heavily investing” in technology that would reduce the amount of power and water required for each AI prompt.
