AI & Environmental Impact: What Happens When You Use Comprehend

When patients or clinicians ask whether using AI tools has an environmental cost, the honest answer is yes, a small one. Here's what we know.

Every time the Comprehend button is used, session notes are sent to an AI model that processes them inside a large computer facility called a data center. Like all computer facilities of this size, data centers consume electricity to run and water to stay cool. This is true of most digital services you use every day, from streaming video to sending email. AI just uses a bit more than most.

How much are we talking?

The most widely cited research (UC Riverside, 2023) that a single AI interaction uses roughly a tenth of a glass of water and about 10 times more electricity than a standard Google search. To put that in perspective, it's still far less than running a dishwasher or driving a mile in a car. The impact is small per use, but across millions of users and interactions, it becomes worth paying attention to.

What's being done about it?

Comprehend PT runs on Amazon Web Services (AWS), one of the largest cloud providers in the world and one that has made significant investments in renewable energy. AWS is actively working toward carbon neutral cloud operations, which means the real world footprint of each interaction is likely lower than older estimates suggest.

What we don't know yet

The AI industry doesn't yet have a standardized way to measure and report environmental impact per query, and per-use environmental data isn't something AI companies currently publish publicly. This is something we've flagged and are actively following as the industry evolves.

The bigger picture

We think about this as part of a broader value equation. Time saved on documentation means less screen time and more time with patients, and that tradeoff matters. We're committed to staying informed as this space evolves and will update this article as better data becomes available.

Updated on: 07/04/2026