The Hidden Cost of ChatGPT: How to Use AI More Sustainably & Ethically

Have you ever thought about what’s happening behind the scenes when prompting ChatGPT? I don’t mean the technical aspect of how it works, but rather how it could be impacting our world from an environmental and ethical perspective. Let me preface this blog by saying, I’m not suggesting you stop using these tools, but recommend being more mindful in how you use them.

When I first started using ChatGPT, I wasn’t struck by “magic”—I was struck by efficiency. As a natural born researcher, I’m used to juggling endless tabs, cross-referencing sources, and chasing context down rabbit holes. With ChatGPT, I could suddenly get a nuanced answer or brainstorm session in seconds, all without the digital clutter and the possibilities felt infinite.

But here’s what most people (my past-self included) don’t realize: every prompt you send to an AI model like ChatGPT has a real-world cost. Not just in dollars or data privacy, but in energy, water, and carbon emissions. Multiply that by millions of prompts a day, and the impact is anything but trivial.

I’m not here to shame anyone—AI is an incredible tool, and I still use it daily. But I’m also a facts-first person, and the facts are worth considering if we want to use this tech responsibly.

So, a single ChatGPT prompt is more than a Google search, but far less than streaming video or sending a big email. The catch: at scale, those grams add up fast.

And then there’s the impact on water. AI data centers rely on water for cooling—Microsoft reportedly used nearly 700 million gallons in 2022, much of it tied to AI activity. (The Register)

The Flip Side: Prompting Can Save Resources, Too

Here’s where my devil’s advocate side kicks in. Yes, AI has an environmental footprint, but so do all the other things we do in business and life. And when used well, AI can actually reduce waste and inefficiency:

• Fewer meetings and emails: A good prompt or AI meeting notetaker (hello, Fathom) can replace a dozen back-and-forths, cutting down on unnecessary digital noise.
• Smarter decisions: AI can help teams quickly analyze data, forecast, and avoid costly mistakes—meaning less waste and fewer do-overs.
• Better automation: AI can streamline processes, eliminate manual data entry, and reduce reliance on paper or inefficient software.
• Improved accessibility: It levels the playing field for non-native English speakers, neurodivergent thinkers, and solo operators.
• Creative iteration: You can prototype ideas faster, without 50 rounds of brainstorming or redundant drafts.

Used with intention, AI can offset its own environmental cost by making other systems leaner and more efficient.

The Ethical Side of AI Use (aka: We Don’t Need More Misinformation)

Environmental impact is only part of the story. We also need to talk about how AI is used—and misused.

• Fake experts and spam content: We’ve all seen the regurgitated blogs, AI-generated headshots, and “thought leaders” parroting chatbot answers. It’s not just lazy—it’s digital pollution.
• Labor and data ethics: Behind the scenes, underpaid workers filter toxic content for model training, and creators’ work is often scraped without consent.
• Bias and misinformation: Prompt carelessly, and you risk amplifying stereotypes or spreading inaccuracies—because AI is only as good as its training data.

If you’re using AI without editing, disclosing, or reflecting, you’re adding to the problem, and we can do better. I literally just shipped out a blog the other day that I used ChatGPT to help create. I tweaked it of course, but after hitting the publish button I realized I hadn’t done my due diligence on the examples it gave. After some quick research, I realized 3 out of the 5 examples were inaccurate. I was kicking myself for weeks. It also made me rethink every blog I’ve read over the past few years and how they could have been completely false. Yikes.

How to Use AI More Sustainably and Ethically

This isn’t about guilt—it’s about being smart and conscious with a tool that can do real good.

For Individuals:

• Batch your thinking and prompts to reduce unnecessary back-and-forth.
• Use lighter models (like GPT-3.5) for routine tasks; save GPT-4 for when you need depth.
• Don’t generate content you won’t use.
• Always edit and add your perspective—AI is a starting point, not the whole answer.
• Disclose when you use AI in public-facing work.

For Teams and Businesses:

• Create internal AI guidelines—not to restrict, but to align.
• Push for authenticity. If your output sounds like a bot, your audience will notice.
• Audit your tech stack and choose sustainable providers where possible.
• Train your team to prompt with purpose, not just curiosity.
• Build workflows where AI actually solves a problem, not just adds steps.

All facts and comparisons referenced are based on industry estimates and published reports as of 2022–2025. For more details, see Microsoft’s and OpenAI’s public sustainability disclosures and independent research on AI’s environmental impact. (Microsoft, OpenAI, MIT, SSRN).

At Prismo Marketing, our comprehensive suite of services covers everything for a standout digital brand. From crafting visually stunning and user-friendly websites to boosting online visibility with results-driven SEO, we ensure discoverability and engagement. Our experts collaborate to create a memorable and cohesive brand image, incorporating creative graphic design that makes a visual impact, from logos to marketing collateral. Additionally, streamline your business processes with seamless HubSpot CRM integration. We’re here to turn your digital aspirations into a reality. Ready to elevate your online presence? Contact us today, and let’s get started on propelling your brand to new heights.