Skills4AI

Raise awareness of the environmental impact of AI

  • 30/04/2026
  • Greta du Velay

Suitable for all audiences, no technical knowledge required.

Every time a student asks an AI a question, sends an image for analysis or generates text, they are consuming energy — often without realising it. These activities, multiplied by millions of users every day, power energy-intensive data centres that emit CO₂ and consume significant amounts of water. AI is not a neutral tool: learning to use it sensibly is also an environmentally friendly choice.

  • A practical introduction to environmental footprints, delivered in a concise format.
  • Accompanying educational resources enable participants to gain a deeper understanding of how AI works, either during the workshop or afterwards.

An extension has been added to the compar:IA interface, providing access to all the workshop content: the workshop plan, suggested prompts, teaching resources and discussion cards.

To run the workshop effectively, it is essential to familiarise yourself with the workshop materials, understand how a debate works, and familiarise yourself with the teaching resources and debate cards before the session.

Objectives

  • Compare the responses from different AI models.
  • Encourage reflection on the environmental impact of the models.
  • Measure the ecological footprint of the questions posed to the AI.

Example of implementation

Here is an example of a session plan for discussing the environmental impact of AI
1 Introduction — 10 min

A brief introduction to generative AI and the compar:IA platform. Accessible to everyone, even without prior knowledge.

2 The duel — 10 to 15 mins

Participants ask the same question to two anonymous models in a ‘David versus Goliath’ scenario (one well-known, larger model and one lesser-known, smaller model). The prompt is chosen according to the group’s needs.

3 Comparison and vote — 10 mins

Group discussion on the suitability of the responses based on participants’ needs. Vote for the best model.

4 Model reveal — 10 mins

Exploring the technical specifications of the models. The answer considered the best is not necessarily the one from the larger model.

5 Final discussion — 10 to 15 minutes

A group discussion to consider the trade-offs between user experience and environmental impact.

6 Conclusion and next steps — 5 mins

Sharing simple steps to reduce the environmental impact of using AI in everyday life.

You may find this resource useful.

All use cases

"Explore our full range of use cases—click below to learn more!"
Click Here