AI avatars and carbon footprint: how to reconcile innovation and responsibility?

Artificial intelligence profoundly transforms the way we create educational content. In immersive training, AI avatars open up new possibilities: presenting information, guiding learners, simulating professional situations or even embodying virtual experts.

But behind these innovations, an increasingly important question arises: what is the environmental impact of these technologies?

At VRAI Learning, we use AI to enrich training experiences in VR and XR. We are convinced of its potential, but we also think that it is essential to take into account the energy footprint of these tools and to make responsible choices. Moreover, there are many requests for support for "CSR education", so let’s be consistent!: )

Generative AI: a powerful technology... but energy-hungry

As we know, the generation of content by artificial intelligence requires computing power. Each image, animation or video generated mobilizes servers and GPUs.

For example, some estimates indicate that generating an image with AI consumes on average a few watt-hours of energy, which remains modest on a small scale but becomes significant when volumes increase.

Recent research also shows that some image generation models can consume up to 46 times more energy than others to produce a comparable result*.

Not all AIs are the same when it comes to energy consumption...

Photorealistic, stylized or video avatars: different impacts

Not all AI avatars require the same level of computation.

Photorealistic avatars seek to reproduce the human face with a very high level of detail: skin texture, hair management, micro-expressions, realistic lighting. This level of precision increases the necessary computing power

Some platforms also offer AI-generated video avatars, where the system produces a complete video from a text. In this case, the AI generates many images to create the animation. A single minute of video can represent thousands of generated images.

Conversely, stylized or semi-realistic avatars generally require fewer resources. They remain expressive and efficient in conveying a message, while being lighter to produce and use.

In immersive environments like VR or XR, this choice can also improve technical performance and the accessibility of experiences.

Our choice: effective and responsible avatars

At VRAI Learning, even if we know how to make photorealistic models, we make a clear choice: prioritize cartoon or semi-realistic avatars rather than totally photorealistic ones.

This choice is motivated by several reasons:

First, these avatars are more suitable for immersive environments like VR or XR. They are more readable, more expressive and often better accepted by users.

But there is also an environmental reason!

Stylized avatars require less computation, fewer graphics resources and less energy than ultra-realistic avatars. In projects that can be deployed with many learners, this difference becomes important.

We are fully aware that AI remains an energy-intensive technology. The objective is not to claim that it is neutral for the environment, but to make choices that limit its impact when possible.

Innovate without forgetting responsibility

The question is not to give up AI. It brings considerable benefits for training: customization, interactivity, accessibility of content, new educational formats.

The real question is rather: how to use these technologies responsibly?

At VRAI Learning, this involves several principles:

-choose technologies adapted to the users avoid useless calculations, prioritize efficient designs rather than technological overbidding and gradually integrate environmental issues into our design choices.

-Innovation must not only be spectacular. It must also be sustainable and relevant.

-And sometimes, a slightly less realistic avatar allows for creating equally engaging experiences... with a more controlled environmental impact 😃

Testez un avatyar!

Source * ARXVI-Cornwell University