Why generate images with AI?
In a context where the required visuals are not always accessible in banks of traditional images, the generation of tailor -made illustrations makes it possible to fill a lack of educational visual content. Whether it is a question of illustrating a historical scene, visualizing a scientific concept or of contextualizing a social problem, these personalized images can enrich pedagogy by arousing interest, by facilitating the understanding of abstract concepts or even illustrating delicate situations that it would be difficult to capture differently.
Here are some ideas offered by participants in this question:
And here are also reasons not to use AI to generate images!
How to generate a quality image?
The workshop has highlighted the importance of formulating the requests addressed to generative AI tools. For example, we can consider AI as having a “visual impairment”, which encourages a most meticulous description from the first request. The request gains to specify in particular the expected format of the image, its style, the visual elements that we want to find there, the action represented as well as a possible feeling to be transmitted.
The demonstration of how the images generated can be modified to serve educational objectives, in particular using tools such as Canva or via publishing functions assisted by AI, was also made during this conference. An exercise on the transformation of a polluted city into an eco -neighborhood specifically illustrated this potential. The question of recent trends that have flooded social networks, such as the famous “starter packs”, and even the generation of humorous memes from existing images, were also discussed.
Here are platforms used by participants for the generation of images:
Ethical, educational and legal issues
After exploring the why and how of the generation of images by IA, the conference has focused in the second part on responsible use of AI. To this end, three main types of issues were discussed:
- The educational criterion: The teacher is invited to question the relevance of the AI for achieving learning objectives, his impact on motivation and understanding, as well as the quality of the AI assisted assessment. To concretely support everything, we proposed to discover a playful tool created by Laurent Di Pasquale of the trendy school, An interactive story which is aimed at teaching staff.
- The ethical criterion: The generation of images has a significant energy cost. Recent studies suggest that generation of a single image can consume the equivalent of a half-load of cell phone. Ethics also involves raising awareness among students, disinformation and transparency on the origin of images. We can summarize this criterion by: “It is not because we can … that we must”.
- The legal criterion: Questions surrounding copyright, the protection of personal information and the use of sensitive content has been raised. In turn, this criterion can be summarized by: “It is not because we can … that we can! »»
These three families of issues are also explained in an accompanying guide Designed by the Quebec Ministry of Education to help teaching staff to think about the use of generative AI in class, taking into account criteria such as quality, inclusion, transparency and agency. He offers a series of reflexive questions to ask before, during and after use.
Moreover, the conference made it possible to ask the participants what issue they consider the most important:
The importance of finding your balance
In conclusion, the participants were invited to question the integration of AI into their practice: what tasks can she lighten? What type of learning does it favor?
They were also able to remember the importance of developing a critical and informed posture in the face of digital technologies in general, and of educating students there now. Without demonizing the uses of AI, the fact that we are also interested in the impacts of this digital activity is very beneficial, while not forgetting to take into account all of our own uses before taking a critical look at those of others. Moreover, several people who followed the wave of “starter packs” or the generation of videos from a simple photograph have received personal attacks on this basis.
Let us not forget that generating an image does not necessarily have more environmental impact than listening to Netflix or scrolling videos on Tiktok without specific purposes. “It is a question of personal values; It is ultimately a question of finding your own balance between the different issues and the perceived added value. It’s complex, but interested in it is a big step in the right direction! »»
The generation of images by AI, if it is well supervised, can become a powerful educational lever, provided that we respect ethical and legal standards and always assess its educational value.
A collaborative wall of resources is available to deepen the subject: padlet.com/ecolebranchee/imagesia
When AI no longer knows!
Here is an extract from a conversation presented at the conference and which shows the ethical limits of an AI at work:
The conversation was able to succeed when the references were removed from Chechnya and the journalist.
To find out more:
