How to Make Your AI Use More Environmentally Friendly

You can't single-handedly change how data centers are built or powered. But the choices you make about how you use AI do add up, especially across millions of users, and a few specific habits make a real difference.

Not Every Question Needs the Biggest Model

One of the most actionable habits is matching the size of the AI model to the size of the task. A short email or a quick rewrite doesn't need the most powerful, most computationally expensive model available; a smaller, more efficient model can usually do the job using meaningfully less energy (Indeed Innovation). One sustainability-focused team that builds AI assistants for clients puts it simply: think of it like choosing a bike instead of a truck for a two-minute errand (Indeed Innovation).

Batch Similar Tasks Together

Every individual request sent to an AI system, whether it's a chatbot or an image generator, requires its own round of processing. Grouping related tasks together, rather than sending many small, separate requests back to back, reduces total processing overhead and the number of round trips to a data center (Indeed Innovation).

Consider Whether You Actually Need AI for the Task

Not every question that gets typed into an AI chatbot needs to go to an AI chatbot at all. For simple factual lookups, a basic web search remains a meaningfully lighter-weight option computationally than a generative AI query, since it doesn't require running a large model to produce an answer from scratch each time. Using AI deliberately, for tasks where it actually adds value, rather than as a reflexive first step for everything, is one of the simplest ways to reduce unnecessary energy use (Girls Who Code).

Run Models Locally When You Can

This is the most structural option on this list, but also one of the most effective: AI tools that run directly on your own device, instead of sending your request to a cloud data center, skip the data center round trip entirely for that task. Sustainability teams researching this specifically recommend running models locally where possible, particularly for frequent or routine AI use, since it reduces the energy-hungry data transfers and cooling demands that come with cloud-based processing (Indeed Innovation).

Choose Providers With Real Clean Energy Commitments

Not all data centers run on the same electricity mix. Some providers have made specific, published commitments to power their operations with renewable energy, or to locate operations where clean energy is locally available. Choosing providers that disclose this clearly, and that back it with measurable progress rather than vague language, rewards companies that are actually investing in cleaner infrastructure, rather than companies that simply talk about it (Indeed Innovation).

Talk About It

This sounds soft compared to the technical habits above, but awareness genuinely shapes behavior at scale. Most people simply haven't encountered the numbers on AI's energy and water footprint, because it isn't usually visible in the product itself. Spreading accurate information, rather than vague alarm or vague reassurance, helps more people make this kind of deliberate choice (Girls Who Code).

The Bottom Line

None of these habits require giving up AI, and none of them will single-handedly offset a global industry's growing electricity and water demands. But used consistently, they shift real demand away from the most energy-intensive patterns of AI use, and they cost you almost nothing to adopt.

Sources

• Indeed Innovation, Using AI Responsibly: 6 Ways to Make Your AI Usage More Sustainable

• Girls Who Code, Eco-Conscious Ways To Use AI