The Hidden Environmental Effects Of AI In Digital Marketing

In 2015, a shaky handheld video shot by PhD student Christine Figgener went viral. 

In it, researchers struggled for nearly ten minutes to pull a plastic straw from the nose of a sea turtle. The turtle’s visible pain sparked outrage around the world. That single video became the rallying cry for the plastic straw revolution.

Almost overnight, “Skip the Straw” campaigns popped up everywhere. A few years later, the movement reached its peak. Starbucks announced plans to phase out plastic straws in favor of sippy-cup lids. Cities like Seattle and San Francisco passed bans. California followed with state-wide restrictions that prevented restaurants from automatically handing out straws. Saying no to a straw suddenly became shorthand for doing your part to save the oceans.

It was a noble cause. A small act that felt meaningful. Unfortunately, it also missed the bigger picture. Plastic straws only account for about 0.025% of the plastic waste in the ocean, while fishing gear, bottles, and packaging account for the overwhelming majority. We zeroed in on a symbol while overlooking the real causes of pollution in our oceans. 

Today, something similar is happening with the conversation around AI. Headlines warn about AI’s massive energy use and carbon footprint, and people are quick to label it an environmental villain. But like plastic straws, are we missing the forest for the trees?

We need to be frank about the environmental impact of AI and machine learning energy use, and about where we should concentrate our efforts to reduce our climate footprint. For digital marketers in particular, it’s a conversation worth having since AI now powers nearly every modern marketing function, from SEO to customer segmentation.

Here’s the question: if you work in digital marketing, is using AI harmful to the environment? 

Let’s find out!

Is AI Harmful For The Environment?

Yes, there’s no doubt about it. 

But so is almost everything about modern life. Our smartphones rely on rare minerals mined in fragile ecosystems. We drive cars and board planes that burn fossil fuels. And we spend most of our days connected to energy- and water-hungry servers powering everything from cloud storage to social feeds.

But when nearly every action has an environmental cost, even the smallest reductions in our footprint matter.  

In the sections that follow, you’ll see comparisons between AI and some of the world’s most polluting industries, like livestock and textiles, to help illustrate scale. But let’s be clear: those sectors should never be the benchmark for what’s acceptable. The goal isn’t to say AI isn’t that bad, but rather to recognize that every emerging technology, especially one as transformative as AI, deserves scrutiny and improvement.

Let’s take an honest look at AI’s environmental impact and explore how we, as digital marketers and technologists, can use it more responsibly to create meaningful change.

How Much Energy Does AI Really Use?

You’ve probably seen the viral claim that a ChatGPT query uses 10 times more energy than a Google search. That statistic, often cited to condemn AI, comes from an early estimate by the Electric Power Research Institute showing that a ChatGPT prompt uses about 2.9 watt-hours, compared to 0.3 watt-hours for a Google search (the equivalent of a 9-watt LED lightbulb being on for 19 and 2 minutes, respectively). 

Newer estimates put that number much closer to that of a Google search. Google estimates a Gemini text prompt uses 0.24 watt-hours of energy. In a recent blog post, Sam Altman stated the average ChatGPT query actually uses 0.34 watt-hours of energy, which is in line with Epoch AI’s estimate of 0.3 watt-hours of energy. By those metrics, AI is no worse than using a search engine. 

But thinking in terms of a single prompt or user isn’t very useful. Whether to brainstorm ad copy or analyze customer data, digital marketers using AI more than likely submit several prompts per day. Multiply that by hundreds of workdays a year and the energy use starts to add up. One hundred prompts per day for 260 days equals roughly 78 kilowatt-hours, or about 0.7% of the energy an average U.S. household uses in a month.

[Curious how your own AI use adds up? Try this AI Impact Calculator. Created by AI consulting firm, Management of the Good, it estimates your energy and emissions footprint from using AI based on your model of choice and number of daily prompts.]

While a single query might seem trivial, whether it’s a Google search or AI prompt, at scale those interactions matter. And electricity is only one part of AI’s total footprint. 

The Bigger Picture: Data Centers, Water, and Emissions

AI systems run on data centers, and data centers run on electricity. And a lot of it. In 2024, data centers worldwide consumed around 415 terawatt-hours (TWh) of electricity, or about 1.5% of global demand, according to the International Energy Agency. That number is expected to grow as more businesses adopt AI and cloud services. 

But this isn’t just about AI. The internet’s environmental toll didn’t start with ChatGPT. For decades, data centers have been guzzling electricity and water to power the websites, ads, analytics, and content that keep the digital economy running. 

Digital marketing has always had a footprint. We just didn’t talk about it much. AI is simply amplifying a problem we’ve been slow to solve. Training and operating large models adds a new layer of energy demand to an infrastructure that was already under strain. 

That puts the responsibility on all of us who build, host, and market on the web. Every online activity, from streaming videos to running programmatic ads, draws from the same shared resources. The energy that fuels AI is also what keeps your website online, your analytics running, and your campaigns reaching customers.

That’s what makes this such an important conversation for marketers to have. The number of prompts we write isn’t as important as examining the entire infrastructure our work relies on. 

For example, streaming one hour of HD video can use more energy than dozens of AI prompts combined. That’s not to let AI off the hook. Instead, it’s a reminder that sustainability in marketing has to consider the whole digital stack, not just the latest tool.

When it comes to data center electricity use, the chemical, paper, metals, and construction industries all consume much more overall energy. But so what? When it comes to climate change we shouldn’t be pointing the finger at other industries. Our focus needs to be on where we can make an impact within our own industry; on the systems we influence. And as a relatively young field, AI has the opportunity and responsibility to grow with sustainability in mind from the start.

What about water? 

In addition to requiring a lot of electricity, data centers also use enormous amounts of water for cooling. In the U.S., they collectively consume about 163.7 billion gallons of water annually; globally, that number rises to around 560 billion gallons. That’s roughly equal to:

• 848,000 Olympic-sized swimming pools
• 10 days of Niagara Falls’ flow
• 0.4% of Lake Erie’s volume
• 1 inch of rain over South Carolina

For comparison’s sake, let’s take a look at arguably the worst industries when it comes to water usage. 

The beef industry, for example, is extremely water-intensive. The average water footprint of beef is 15400 liters/kg or 1,846 gallons per pound. Considering the average American consumes 67 pounds of beef, that translates to 123,682 gallons of water to meet each American’s beef consumption. With a population of 340.1 million people, 42 trillion gallons of water are used each year just to meet US demand for beef.

Meanwhile, the fashion industry is even thirstier at a staggering 56.78 trillion gallons. That’s around 86 million Olympic swimming pools or roughly 44% of Lake Erie’s volume. 

While AI’s footprint is still far less than those water-intensive industries, these comparisons shouldn’t make us comfortable. They’re a reason to proceed with caution. We’ve seen what happens when industries grow first and plan for sustainability later. It’s on us to make sure AI doesn’t follow that path. 

There’s also the question of where that water comes from. Most data centers draw from municipal supplies, the same potable water that communities depend on for drinking, sanitation, and agriculture. During droughts or in regions already facing water stress, that dependence can create real ethical tension.

There’s also the issue of what happens after. Cooling systems don’t return water in the same condition they take it. Once used, that water often carries chemical additives (to prevent corrosion or bacterial growth) and thermal pollution, meaning it re-enters waterways at higher temperatures that can harm local ecosystems.

As workers who actively rely on these technologies, we need to demand accountability from technology providers. We need to work together to ensure AI and the digital infrastructure it relies on evolve differently, using recycled or non-potable water sources, investing in closed-loop cooling, and prioritizing transparency about their environmental impact.

Global Emissions: Context and Opportunity

Data centers currently contribute about 1% of global greenhouse gas emissions, according to the IEA, while the fashion industry accounts for 10% and livestock agriculture for 15%. But what matters now is the rate of change.

The IEA projects that global electricity consumption from data centers will more than double by 2030, rising from approximately 415 TWk in 2024 to about 945 TWh. As for AI, the energy use of accelerated servers (driven by AI workloads) is projected to grow by nearly 30% per year in the baseline scenario. 

In the U.S., a separate report suggests data centers’ share of total electricity could rise from around 4.4% in 2023 to between 6.7% and 12% by 2028. Another projection from Goldman Sachs forecasts that global power demand from data centers could increase by as much as 165% due to AI by the end of the decade, compared to 2023.

These numbers paint a clear picture. AI’s footprint is still smaller than big polluter industries, but the growth trajectory is steep. If left unchecked, the jump in energy consumption and emissions could significantly erode the small footprint argument. 

That said, AI also has the potential to reduce emissions through smarter systems. According to a report by the Grantham Research Institute on Climate Change and the Environment, advancements in AI across energy, transport, and agriculture could cut 3.2 to 5.4 billion tonnes of CO₂-equivalent annually by 2035. Similarly, Google was able to reduce the energy required to cool its data centers by 40% thanks to DeepMind.

That means the same technology powering your ad strategy could also help build more sustainable businesses.

For example:

• AI can model energy consumption trends to help brands meet ESG goals.
• Machine learning can optimize supply chains to reduce shipping waste.
• Predictive analytics can guide sustainable consumer behavior campaigns.

When we view AI through this lens, it becomes less of a threat and more of a tool for impact. It becomes a way to build eco-friendly marketing technology that benefits both people and the planet.

The Environmental Impact of AI Will Lessen With Time

Fortunately, it’s not all doom and gloom when it comes to data centers. History shows that as technology evolves, efficiency wins. For example, early incandescent lightbulbs wasted 90% of their energy as heat. Today’s LEDs use 75% less energy and last up to 25 times longer. Cars have cut harmful emissions by over 70% since the 1970s and get more than double the miles per gallon. And air travel, once a carbon-heavy luxury, is now roughly 80% more fuel-efficient than the first commercial jets. 

We can expect AI to follow the same trajectory. Researchers like David Patterson at Google have shown that better model design, hardware, and energy sources can cut AI’s energy use by up to 100× and emissions by as much as 1,000× for equivalent workloads.

Other researchers echo similar findings. A 2024 MDPI review on energy-aware machine learning found a growing push toward “green AI,” while Bolón-Canedo et al. (2024) noted the same shift: researchers and engineers are prioritizing eco-friendly architectures and energy-efficient training methods across the board. 

AI tools can help reduce waste, too. Smarter audience targeting means fewer irrelevant impressions. AI-powered optimization ensures campaigns run at the right times and places, minimizing data churn. And predictive modeling can help brands forecast demand more accurately, reducing overproduction and excess inventory, especially in retail and e-commerce.

That’s a reason for optimism for us digital marketers. The tools we rely on today are becoming more efficient behind the scenes. But that doesn’t mean we get to stop thinking about our impact. We still have choices to make about how we use those tools. 

Sustainable AI in Digital Marketing

Here’s where AI and marketing meet sustainability head-on. If you want to reduce your carbon footprint while keeping your digital strategy sharp, start here:

1. Use green hosting and cloud services

Choose providers that run on renewable energy and publish transparent sustainability reports.

2. Batch and optimize your AI use

Instead of generating assets piecemeal, group your AI requests. Run prompts in batches, use smaller models for lightweight tasks, and avoid unnecessary prompting.

3. Streamline your creative assets

Video and image files drive far higher energy costs than text-based AI tools. Compress media, avoid auto-play video ads, and use adaptive image formats to reduce data load.

4. Audit your tech stack

Evaluate your marketing automation, CRM, and analytics platforms. Opt for vendors that disclose their energy and data practices, or partner with agencies that hold them accountable.

5. Educate and offset

If your brand tracks sustainability metrics, include digital activity in your emissions reporting. Offset through verified programs like Gold Standard or the Climate Label, and communicate your efforts transparently.

6. Lead with values

Make sustainability a visible part of your brand story. Ethical marketing isn’t just about messaging; it’s about operational choices that align with your mission.

By integrating these habits, your marketing can be both data-informed and environmentally conscious. It’s a win-win for your business and the planet, and that’s definitely something we can get behind. 

Related: How Sustainable Practices Enhance Brand Loyalty

So, is AI bad for the environment?

The short answer is yes, along with all the other technologies we use as digital marketers. 

The more nuanced answer is that while the current emissions from AI-powered infrastructure may feel manageable, the window to act is now. Left unchecked, growth will amplify the environmental impact fast. But we also need to consider the entire data center landscape instead of just zeroing in on AI. In other words, we can’t get caught up on banning the plastic straw. 

The opportunity for marketers right now is to lead the charge toward a more sustainable digital ecosystem. By using AI responsibly, optimizing digital assets, and demanding transparency from tech partners, we can make our corner of the internet greener, cleaner, and more human-centered.

Every click, search, and prompt leaves a footprint. But together, we can make sure it’s a lighter one.