Should Sustainability Teams Be Worried About the Footprint of AI?

Should Sustainability Teams Be Worried About the Footprint of AI?

AI is booming—and so are the questions about its environmental impact. For sustainability-minded companies, the concern is real: Is using tools like ChatGPT or internal copilots irresponsible from a climate perspective? Should AI use be avoided, limited, or even banned?

The short answer: Not necessarily. Yes, AI uses more energy than traditional software—but it’s impact relative to the rest of your business may not be substantial . Let’s unpack why.

‍

What We Know About AI’s Energy Use

The AI industry is increasing energy consumption at a very rapid pace and it is important to note your organization’s contribution to that increased demand. It has been estimated that a single ChatGPT query uses roughly 10 times more energy than a Google Search. That sounds dramatic—but we’re still talking about millisecond-scale, single-digit watt usage per query. This isn’t equivalent to manufacturing steel or flying a plane. It’s more like the jump from using a basic app to running a video call: higher intensity, but within a familiar range.

AI may consume more energy per use than traditional software, but that doesn’t mean it’s inherently unsustainable. 

Context Is Everything

Let’s take a step back and compare emissions across a few example categories.

‍

‍

In other words, using AI is orders of magnitude less carbon-intensive than most physical goods, and similar in scale to many other digital operations we don’t think twice about. 

Accounting for AI use in your Corporate Inventory

The rise of AI parallels the early days of cloud computing. When the cloud took off, it raised similar concerns—energy, data center emissions, and opaque value chains. But the industry matured. Measurement evolved from spend-based  emissions factors (EFs) toward supplier-specific data.

That’s where AI is likely headed, too.

Today, AI energy consumption is difficult to measure and we estimate AI emissions using spend-based EFs (typically using the EFs we already use for software), but we will likely see companies start to become more transparent as their customers demand it, much like cloud service providers have worked to disclose more granular emissions data (albeit imperfectly). As more AI providers release better energy consumptiondata we’ll get better and more accurate baselines.

In the meantime, we expect the industry to continue to use spend based approaches to quantify AI emissions. This means that most inventories will only capture the AI usage that is paid for (typically through subscriptions), and free versions will go unquantified just like google searches do today.

What Sustainability Teams Should Do Now

1. Avoid panic. Unless you are an AI-based software company, the use of AI doesn’t radically break your footprint.
   
   
2. Use spend-based estimates to track usage today. These are already baked into many LCA and carbon accounting platforms.
   
   
3. Push vendors for transparency. Just like we ask suppliers to share LCAs, emissions, and energy consumption, we can do the same for AI tools.
   
   
4. Notice the upside. AI is already helping optimize shipping routes, conduct research and analysis faster, and accelerate R&D—applications that can lead to measurable emissions reductions.
5. Be wise and mindful of AI use. As with everything, there are ways to utilize AI tools to reduce the overall energy for a given task. There are also frivolous uses that drain resources without adding value. 

In Summary

AI uses energy. But that’s not a reason to reject it outright. After all, the emissions from one AI-assisted report aren’t going to rival your packaging, product sourcing, or transportation footprint anytime soon.

But be wise about your use of AI. Know that it does take resources to operate and significantly more than a traditional internet search. But find ways that the AI drastically reduces your task time. Using it for tasks that it is more efficient at than you, such as researching, analyzing data, or report writing, can result in a reduction in total energy use. 

If you use AI to help you draft a report that lets you close your laptop an hour early and get out into nature, that’s a win for you and for the environment.

If you are a sustainability professional concerned about your company’s use of AI, Planet FWD can conduct a free assessment to put the estimated emissions from AI in the context of your company’s total emissions. Request that free assessment here.