Artificial Intelligence for Sustainability

Generative AI Use for Sustainability:
A Zero-Sum Game?

In a new study, almost half of execs admit their use of gen AI has fueled their GHG emissions and 42% have had to re-examine their climate goals — because too many still have a one-eyed view of ‘costs.’

Published 59 minutes ago | About a 6 minute read

Image: Brett Sayles

A new report from the Capgemini Research Institute shows explosive growth in corporate adoption of generative AI (gen AI). But many organizations are failing to appropriately track the technology's significant and growing environmental impact, which is jeopardizing their sustainability objectives. As businesses weigh up Gen AI’s ability to drive business growth against the technology’s environmental cost, the report outlines measures to design a responsible and sustainable generative AI strategy.

The double-edged sword of generative AI

Developing Sustainable Gen AI is based on findings from a global survey in August 2024 of 2,000 executives from organizations with more than $1 billion in annual revenue across 12 sectors and 15 countries in North America, Europe and Asia-Pacific — respondents were reasonably/well/highly informed about their organizations’ strategies and initiatives around gen AI, as well as environmental and social sustainability initiatives. The report found that generative AI adoption has accelerated rapidly: Previous Capgemini research showed that 6 percent of organizations had integrated generative AI across their business functions and locations as of the end of 2023; by October 2024, that figure had risen to 24 percent. AI can help drive sustainable business growth — organizations are leveraging it to help with everything from eliminating greenwash from their communications and food waste from their operations to creating ethical, transparent supply chains and getting a leg up on other common sustainability challenges. But generative AI requires processing vast amounts of data and significant computational power — which consumes massive quantities of electricity, water and other resources — and the limited lifespan of gen AI hardware will exacerbate the tech industry’s e-waste problem: An October 2024 study predicted a 1,000-fold increase in gen AI-related e-waste by 2030.

According to the report, nearly half (48 percent) of executives admit that their adoption of generative AI has fueled their company’s GHG emissions — an increase that is expected to continue to grow: Organizations that currently measure their gen AI footprint expect its contribution to their total carbon emissions to rise, on average, from 2.6 percent to 4.8 percent over the next two years. To mitigate this, organizations are increasingly turning to renewable energy sources and optimizing their AI infrastructure.

Few organizations considering sustainability implications of gen AI

Organizations’ sustainability reporting has not kept up with the rapid pace of innovations around gen AI. Only 12 percent of execs that use gen AI say their organization measures the environmental footprint of its use, and only 38 percent claim to be aware of that environmental impact. Similarly, performance, scalability and cost are key considerations for gen AI model evaluation; sustainability is only of marginal importance. While more than half of execs recognize that making sustainability a key criterion in vendor selection for generative AI-related requirements would reduce environmental footprint, only one fifth rank the environmental footprint of Gen AI as a top-five factor when selecting or building Gen AI models.

But while financial costs may be top of mind for selecting or building gen AI models, the report points out the importance of factoring in how the resource costs of these models affect the bottom line as well.

Industrywide support needed to properly account for gen AI’s impacts

With a growing awareness of gen AI’s environmental impacts, almost a third (31 percent) of organizations have taken steps to incorporate sustainability measures into the Gen AI lifecycle; and over half are either already using smaller models and powering gen AI infrastructure with renewable energy sources or planning to do so in the next 12 months.

However, with more than three-fourths of organizations using only pre-trained models and just 4 percent building their own models from scratch, executives are heavily reliant on their technology partners when it comes to addressing gen AI’s environmental footprint. In fact, nearly three-quarters find it challenging to measure the technology’s footprint due to limited data/transparency from providers and the industry lacks a methodology around how to account for its environmental footprint.

As Niklas Sundberg, Chief Digital Officer and SVP at global transport and logistics company Kuehne+Nagel, has pointed out: “You should be able to ask Copilot or ChatGPT what the carbon footprint of your last query is, but none of the tools will give you a response to that question at the moment.”

In its 2024 environmental report, Google revealed that its emissions had increased by 48 percent in four years thanks to the expansion of its data centers to support increased AI usage. Because of this, the company has admitted its goal to reach net-zero emissions across its entire operations and value chain by 2030 is now “extremely ambitious” and “will require (Google) to navigate significant uncertainty.”

As Google stated in the report: “While we remain optimistic about AI’s potential to drive positive change, we’re also clear-eyed about its potential environmental impact and the collaborative effort required to navigate this evolving landscape. We’re committed to responsibly managing the environmental impact of AI by deploying three major strategies: model optimization, efficient infrastructure and emissions reductions.”

A roadmap for sustainable, responsible Gen AI use

“If we want gen AI to be a force for sustainable business value, there needs to be a market discussion around data collaboration — drawing up industry-wide standards around how we account for the environmental footprint of AI — so business leaders are equipped to make more informed, responsible business decisions and mitigate these impacts,” said Cyril Garcia, Capgemini’s Head of Global Sustainability Services and Corporate Responsibility and Group Executive Board Member. “We are proposing here practical steps to follow for business leaders to fully harness technologies such as gen AI and deliver a positive impact for organizations, society and the planet.”

Echoing SustainableIT.org’s Responsible AI Framework, the report suggests that businesses:

conduct a thorough assessment of both the financial ROI and environmental footprint of generative AI projects before launch — ROI calculations of gen AI applications should consider resource costs as a relevant factor.
consider whether they need generative AI in certain cases or whether less resource-intensive technologies would suffice — in addition to tapping smaller, fine-tuned models; using a multiple-model approach could offer a range of latency, accuracy and carbon footprint solutions. Using agentic AI could also optimize cost as well as lower energy use.
make smart choices and implement sustainable practices throughout AI’s lifecycle — including hardware, model architecture, energy sources for data centers, and sustainable usage policies.

The report stresses the importance of continued research and monitoring to fully understand and mitigate generative AI’s environmental impact. Multidisciplinary governance models, effective policies and industry-wide collaboration between stakeholders across the gen AI ecosystem will also be important for organizations that want to achieve safe, transparent, sustainable and ethical generative AI usage.

Organizations that manage to innovate and control the cost and energy usage of gen AI will be better positioned for success in the long run. Cost and carbon impact go hand in hand; addressing both will drive innovation, improve economics and help make gen AI a truly sustainable tool.