In short
Teach a handful of habits in a short session with real examples from your own work: check whether AI is the right tool, write clear and specific prompts, ask for short answers, refine rather than regenerate, and go easy on images and video. Back the training with efficient defaults in company tools, because defaults reach people who skip the course.
Steps
- Pick three or four real tasks staff already use AI for and use them as the examples.
- Teach five habits with a before-and-after prompt for each.
- Explain in one slide why output length, images and regeneration drive energy use.
- Set efficient defaults in the tools staff use so the habits do not depend on memory.
- Add the session to induction and refresh it when tools change.
Short, example-led training plus good defaults works better than a long lecture on energy. The same habits save staff time and money, which is usually the easier case to make.
Which habits should I teach?
- Right tool first. Search, templates, spreadsheets and existing software often do the job.
- Clear, specific prompts. Say what you want, for whom and how long. Paste the relevant section, not the whole document.
- Ask for short answers. UNESCO and University College London found shorter prompts and responses could cut energy use by more than half.
- Refine, don't regenerate. Tell the tool what to change instead of pressing "try again" repeatedly.
- Go easy on images and video. In tests by Luccioni, Jernite and Strubell, generating images averaged about 2.9 kWh per 1,000 images against 0.002 kWh per 1,000 text classifications.
Why does length matter?
Energy rises with how much a model reads and writes. Epoch AI estimates a typical ChatGPT query at about 0.3 Wh, rising to about 2.5 Wh with a 10,000-token input and almost 40 Wh with a 100,000-token input. Pasting a long report into every chat adds up.
What makes the biggest difference
Defaults. If the company assistant starts on a smaller model, limits output length and has unused AI features switched off, everyone benefits, including people who never attend the session.
Common mistakes
- Making it about guilt. Frame it as doing the job well.
- Abstract examples. Use your own tasks.
- One-off training. Tools change; refresh it.
More habits are in use AI efficiently.
Sources
- Smarter, smaller, stronger: resource-efficient generative AI and the future of digital transformation, UNESCO and University College London (2025)
- Power Hungry Processing: Watts Driving the Cost of AI Deployment?, Luccioni, Jernite and Strubell, ACM FAccT (2024)
- How much energy does ChatGPT use?, Epoch AI (2025)
Related
Use AI efficiently: everyday habits
Everyday habits that cut the energy of using AI: clear prompts, fewer regeneration loops, care with images and video, caching and turning off defaults.
How do I write prompts that use less energy?
How to write AI prompts that need less computation: be specific, set a length, paste only what matters and get it right first time instead of retrying.
How do I write a green AI policy?
What a short green AI policy should say: when to use AI, which defaults to set, what to ask suppliers, how to measure, and who owns it.
How do I stop regenerating AI answers over and over?
Why repeatedly pressing regenerate wastes energy and time, and how to get a usable AI answer with one good prompt and a short follow-up instead.