Methods to assess the environmental impact of AI tools in the NHS

What is the environmental impact of using AI tools like ambient voice technology in the NHS?

Artificial intelligence (AI) tools are becoming central to the NHS’s analogue-to-digital transformation. One example is ambient voice technology, which listens during medical appointments and automatically generates notes and letters. This technology has the potential to save time, reduce administrative burden, and ease pressure on doctors and nurses.

But behind the scenes, AI tools also rely on powerful computers, data centres, and internet infrastructure. All of these use a lot of electricity, need large amounts of water for cooling, and contribute to carbon emissions (gases that contribute to climate change).

The NHS has the ambition to be “net zero” by 2045: reducing carbon emissions it can influence and balancing out the rest. To meet this ambition, the NHS needs better ways to understand the environmental impact of new AI tools before they are adopted more widely.

This project will look at how ambient voice technology is currently used in hospital and community settings across the NHS and estimate the carbon emissions linked to it. We will also ask NHS professionals and members of the public what they think about the balance between the benefits and environmental impacts of using AI in healthcare.

Our findings will help NHS organisations to make greener and better-informed choices when bringing in new technologies to improve care.

How will we develop methods to assess the environmental impacts of using AI tools in the NHS?

We will use ambient voice technology as a case study, and:

1. Evaluate how well we can estimate the environmental impact of ambient voice technology in NHS settings using a tool called Green Algorithms.
2. Speak with clinical, digital, and procurement leads about what to measure and how to feed this back when trying to understand the environmental impact of AI tools.
3. Host workshops to explore how patients and the public weigh the benefits of such tools against their environmental impacts.

Our findings will inform new ways of tracking the carbon emissions of AI tools such as ambient voice technology. This will help NHS organisations to make greener choices in procurement and use of AI tools, so that digital transformation remains aligned with the NHS’s commitment to becoming net zero.

Project team

Dr Jan van der Scheer

THIS Institute

LL

Loïc Lannelongue

University of Cambridge

JS

James Smith

University of Cambridge

PB

Philippa Brice

Cambridgeshire and Peterborough Integrated Care System

FB

Fanny Burrows

NHS England

Professor Mary Dixon-Woods

THIS Institute