Starting next fall, a course on AI will be offered to all students in Block 3, with the aim of showing how it was created and how it works, going beyond the "coffee shop" considerations that we hear all too often in the media. The legal, ethical, managerial, and educational aspects of AI will be addressed by teachers from different faculties. We will show how AI is used in each discipline, so that students can gain an overview of both the risks and opportunities in their field. The teaching will be research-based, reminding students that AI is not the preserve of the private sector.
In the medical field, AI is also increasingly being used to aid decision-making. Even if medicine remains an art... If you are 27 years old and have just finished your medical studies, how would you react if AI suggested a diagnosis that you didn't think was correct? If you did what you thought was right and got it wrong, you would blame yourself for not following AI's recommendations. But conversely, if you follow the AI and you were actually right, you'll blame yourself for putting your critical thinking skills on the back burner... Because making good use of AI also requires an understanding of the workings of human psychology... Let's be honest: if AI allows you to make 100 correct decisions, on the 101st, you're not going to pay attention anymore. By shedding light on the technical aspects of AI, this course will provide a foundation from which students can develop critical thinking skills in relation to the "tensions and uncertainties" specific to their discipline.
The use of AI itself is in tension with another major social issue: sustainable development. AI consumes a lot of energy. Yet AI's share of global energy consumption continues to increase... Students need to be made aware of this: if I can solve my problem without AI, it's better from an energy perspective. Twenty ChatGPT queries consume one liter of water. Using AI is a bit like turning on a tap: perhaps we should see it as a common good that should not be abused...
Benoît Frenay, Faculty of Computer Science, NaDI Institute