You don’t need to be told that tools like ChatGPT are now part of everyday working life. AI-generated stories are all over LinkedIn, and you’ve probably seen the “action figure” hype too. And that makes sense. AI tools deliver quick results, inspire, effortlessly rewrite entire texts, and also offer plenty of fun.
The possibilities are impressive, and the threshold for using them is low. But beneath the surface lies a major drawback: the powerful AI models that help you quickly translate that one email run on energy-intensive infrastructure. So every generated sentence requires a lot of computing power, and that has a direct impact on the environment. At the same time, we understand that choosing not to use AI at all means falling behind. Fortunately, there are ways to use AI more consciously and sustainably.
What does one prompt cost?
Let’s first look at what’s really going on. Every time you enter a prompt into a tool like ChatGPT, a sophisticated language model is activated. So every command you give (“translate this email”, “write a blog about AI and the environment”) sets off a lot behind the scenes. The language model processes the input, calculates billions of parameters, and generates a response based on that. It may sound abstract, but in practice, it means: a lot of computing work, done by heavy-duty servers in large data centers. These data centers run on electricity, often supplied by fossil fuels.
According to a study by Hugging Face and Carnegie Mellon University (2022), generating 1,000 responses with a large language model like GPT-3 can consume up to 500 times more energy than a simple Google search. And that’s just for usage. Training such a model costs much more: thousands of GPUs run non-stop for weeks, producing CO₂ emissions that can reach dozens of tons. These GPUs are specialized processors that can process enormous amounts of data simultaneously, but they also use a lot of electricity. That consumption is equivalent to multiple intercontinental flights. So next to flight shame, you might also experience prompt shame.
Repeating prompts increases the impact
Another factor that increases the impact is how we use these AI tools. You often don’t get the perfect text on the first try. You generate several versions, adjust the tone, ask for more variation or simplification. Out of frustration, you might type after the third attempt: “rewrite, but better.” Each prompt restarts the entire computation process. That makes generative AI use iterative and energy-intensive.
Every generated text adds to the growing energy demand. If you’re writing texts that mainly need to meet language requirements like readability, inclusivity, or tone of voice, there are more efficient and less resource-intensive alternatives, such as working from your own text with targeted support, instead of regenerating it every time.
Less computing power, more efficiency: how Textmetrics is different
So how can you still use AI without unnecessarily burdening the environment? Textmetrics offers a lighter alternative. Instead of generating new texts each time, the software analyzes what’s already written. Your existing text is assessed based on clear criteria such as reading level, inclusivity, and clarity. You get instant feedback, allowing you to make targeted improvements yourself—without an energy-hungry language model having to recalculate everything in the background.
For analysis, Textmetrics uses smart linguistic rules, natural language processing (NLP), and, where needed, lightweight machine learning. So no billions of parameters or heavy neural networks, but efficient technology that does exactly what’s needed, with a much smaller ecological footprint.
Concrete benefits at a glance
- Less iteration needed: you don’t need to regenerate your text five times, because you immediately see where and why it can be improved
- Lighter technology: Textmetrics runs on more efficient algorithms than large language models
- Real-time optimization: you work with your own text, not a generated alternative
Use AI where it makes sense, not where it burdens unnecessarily
This is not a plea against the use of generative AI. In many cases, it’s a powerful tool with unique capabilities. But it is important to be aware of the impact. Every prompt brings an energy cost. And the more often you repeat, the higher that cost becomes.
If you optimize texts daily for inclusive or clear communication, you can choose a solution that makes your work lighter. For yourself, and for the environment.