A new study—the latest in a growing body of research—highlights how environmentally-minded lab managers might find it difficult to reconcile AI usage with their lab sustainability goals, particularly in regards to CO2 emissions.
According to the study published in Frontiers in Communication, larger and more accurate large language models (LLMs) achieve their results with “steep increases in emissions” compared to smaller, but less accurate, models. According to the study, one of the largest and most accurate LLMs emitted 50 times more CO2 than one of the smallest (and least accurate) models.
Lab Management Certificate
The Lab Management certificate is more than training—it’s a professional advantage.
Gain critical skills and IACET-approved CEUs that make a measurable difference.
Labs increasingly adopting AI
Since the initial release of ChatGPT in November 2022, scientists have begun using chatbots for a wide variety of tasks, including summarizing scientific literature, writing capital requests, and programming computers. According to one open-access paper, researchers estimated that at least 10 percent of biomedical abstracts submitted in 2024 were generated with the assistance of a chatbot, with that statistic rising to 30 percent for some subgroups across disciplines, journals, and countries.
Similarly, many lab software vendors have incorporated LLM-powered features into their products to assist with inventory management, experiment design, data analysis, and more.
Suffice it to say, LLMs are finding traction in the lab.
But for those mindful of their lab’s environmental footprint, using LLMs will run counter to such concerns. Scope 3 emissions, which are those generated by external sources such as transportation and waste disposal, would also encompass the emissions generated by the datacenters powering the AI models.
Interested in lab design?
Is the form not loading? If you use an ad blocker or browser privacy features, try turning them off and refresh the page.
By completing this form, you agree to receive news updates and relevant promotional content from Lab Design News. You may unsubscribe at any time. View our Privacy Policy
Furthermore, the demands of scientific research would likely lead most labs to only use the largest—and most energy-intensive—models so as not to affect the quality of their work.
The intense energy use associated with LLMs will grow alongside adoption until more energy-efficient models are developed.
Partnering with like-minded vendors for AI services
Lab managers who wish to minimize the carbon impact of their AI usage can partner with like-minded service providers. Here’s a breakdown of the sustainability initiatives of four major LLM service providers: Google, Microsoft, Anthropic, and OpenAI.
- Google, which offers the Gemini LLM chatbot, has a goal of reaching net zero emissions by 2030. While the costs of running Gemini and other projects have complicated this goal, the company is still working toward it by investing in other AI solutions that will ideally “help individuals, cities, and other partners in their efforts to collectively reduce one gigaton of their carbon equivalent emissions annually by 2030.”
- Anthropic, the company behind the Claude LLM chatbot, is famous for its focus on safety and aligning AI with human values. While there is very little information about the company’s sustainability considerations, Lab Manager did find the details to one commitment in their paper introducing the new Claude 3 family of LLMs:
Anthropic works to fully offset our operational carbon emissions each year, partnering with external experts to conduct a rigorous analysis of our company-wide carbon footprint. Once measured, we invest in verified carbon credits to fully offset our annual footprint. Our credits directly fund emissions reduction projects.
The document goes on to state that their goal is to maintain net-zero climate impact through these offsets.
- OpenAI, which offers ChatGPT, has not released any official statements or documents related to net zero initiatives, according to Lab Manager’s research.
- Microsoft’s Copilot, which uses a modified version of the same LLM powering ChatGPT, was developed in the midst of the company’s ambitious plans to become carbon-negative (removing more carbon than it emits) by 2030. But in February 2025, the company acknowledged that the goal was a “moonshot,” especially in the face of AI. That said, Microsoft is willing to lose the battle to win the war. According to Melanie Nakagawa, Microsoft’s chief sustainability officer, AI will “create distance from our [sustainability] goals in the short term” but will help them reach them in the long term.
Many business factors figure into AI decisions. However, looking at things strictly through an environmental lens, sustainability-minded lab managers may be best served by working with Google or Microsoft. While large, accurate AI models offer undeniable benefits, their substantial carbon footprints complicate sustainability targets. Strategic partnerships can help labs mitigate this impact. Ultimately, lab managers should be thoughtful in how they integrate AI into their operations to find the right balance between increased productivity and any environmental goals the lab might have.
