How to Use AI Responsibly (Without Slowing Your Team Down)
AI makes us faster. It makes our work better. Nobody's giving that up, and we're not suggesting you should.
But there's a side of AI that doesn't come up much in marketing circles, and it's worth talking about. The tools we use every day are running on massive data centers that consume enormous amounts of energy and water. And the scale of that consumption is growing faster than most people realize.
This isn't about using AI less. It's about using it smarter.
The Environmental Impact of AI (The Numbers Are Bigger Than You Think)
We did an internal knowledge share on this recently, and some of these stats genuinely caught us off guard.
- Data center energy demand more than doubled between 2017 and 2023, and it's still accelerating. (MIT News)
- A single generative AI query uses about 5x more electricity than a basic Google search. (Girls Who Code)
- ChatGPT alone produces over 573,000 pounds of CO2 every month, roughly the carbon equivalent of 260 flights from New York to London. And by 2030, data center power demand is projected to grow 160%. (Earth Day)
Image generation is in a different category entirely. Research shows that generating an AI image uses roughly 60x more energy than a text prompt. That adds up fast when you're iterating through five versions trying to land on the right one.
These numbers aren't meant to scare anyone away from using AI. They're meant to make the case for using it with intention.
Why This Matters More for Marketing Agencies
Most teams use AI occasionally, while marketing agencies use it all day, across every client, every deliverable, every workflow. That puts us in a different position than the average user.
The habits we build around these tools tend to stick. And if we're being intentional about how we run campaigns, build strategies, and show up for our clients, it makes sense to be intentional here too.
The good news is that using AI more responsibly and using it more effectively are basically the same thing. Clearer prompts get better results. The right tool for the right task means fewer rounds of back-and-forth. Smarter workflows save time and energy, in every sense of the word.
How to Use AI Responsibly: 5 Practical Habits
The habits that make the biggest difference are usually the simplest ones.
- Ask yourself if AI is actually the right tool: Not every task needs a generative AI tool. Sometimes a quick Google search gets you there faster with a fraction of the energy cost. Before you open a new chat, take five seconds to check whether AI is the best path or just the most familiar one.
- Write a clear prompt before you start, not as you go: Wandering prompts waste time and energy. The more specific your prompt is upfront, the fewer rounds you'll need. Front-load your context. Paste in your research instead of asking AI to find it. Show the AI an example of what you want rather than trying to describe it from scratch.
- Match the model to the task: Not everything needs the most advanced version of a tool. Lighter tasks can run on lighter models. Save the heavy-duty processing for work that actually calls for it.
- Be more intentional with image generation: Because the energy cost is so much higher than a text prompt, it's worth pausing before you start generating. Have a clear direction before your first attempt. Write a detailed prompt that covers style, mood, and composition so you have a real shot at getting what you need in one or two rounds, not six. And save assets that work well so you're not regenerating something you've already made.
- Build team habits so nobody starts from scratch every time: When everyone on your team is re-explaining the same client context every time they open a new chat, that's wasted effort on every level. Using Projects in ChatGPT and Claude, building shared prompt templates, and documenting what works means the whole team benefits, not just the person who figured it out.
The RED66 Low-Impact AI Framework
After the knowledge share, we put together a simple internal framework. Think of it less as a rulebook and more as a quick gut-check before you prompt.
Before a single prompt, ask yourself
- Could I find this with a quick Google search instead?
- Am I using the right tool for this, or just the one I always reach for?
- Is my prompt clear enough to get a usable answer without a lot of back-and-forth?
Before an AI-heavy project, ask yourself
- Do I have a clear goal before I start prompting?
- Can I give the AI more context upfront to cut down on the rounds I'll need?
- Am I using a more advanced model than this task actually requires?
- Would pasting in a document or existing research get me to a better result faster?
And when it comes to picking the right tool:
| ChatGPT | is the all-around creative generalist — great for brainstorming campaign concepts, writing copy variations, generating images for mockups, and day-to-day tasks. If you're not sure which tool to reach for, ChatGPT is usually a safe starting point. | Brainstorming, creative copy, image generation, all-purpose tasks | You just need a quick factual answer |
| Claude | tends to nail writing style and tone better than the others — especially when you feed it examples of your best work or a client's brand voice. It's a strong choice for longer content like blogs, brand narratives, or anything where consistency really matters. It's also where the Projects feature really shines for client work. | Long-form writing, editing, brand voice consistency, document analysis, projects with a lot of context | You need real-time or current info |
| Gemini | is the pick if your work lives in Google's ecosystem — think pulling from Google Drive docs, working across Gmail, or handling tasks that involve both text and visuals. It removes the copy-paste back-and-forth of loading context into a separate tool. | Research with Google Workspace, analyzing docs in Drive, multimodal tasks (text + image + video) | You're working outside Google's ecosystem |
| Perplexity | is built for research — it pulls real-time info from the web with cited sources, making it ideal for things like competitor research, industry trend lookups, or fact-checking before a client presentation. It's not a content generator though — use it to gather, not create. | Fact-checking, current events, competitor research, anything where you need cited sources | Creative writing or content generation |
Using the right tool for the job means fewer prompts, better results, and less wasted energy. That's a win across the board.
The Bottom Line
AI is already part of how we work, and being more thoughtful about how we use it just makes everything run better.
The fundamentals of good AI usage aren't complicated. Clear prompts, the right tool for the task, a quick pause before you generate that fifth image variation. Small habits, real impact.
If you're thinking about how your team uses AI and want to build smarter habits around it, we're happy to talk.
