Un MOOC pour initier a l'IA : « Intelligence Artificielle avec Intelligence »

Frédéric Alexandre, Marie-Hélène Comte, Aurélie Lagarrigue, Chloé Mercier, Axel Palaude, Margarida Roméro, Thierry ViévilleSPIRE 2023 · OpenAlex

Abstract

This paper presents the design and impact of "Intelligence Artificielle avec Intelligence" (IAI), a hybrid and participatory MOOC designed to demystify AI for the general public. Reaching over 32,000 participants with a 94% satisfaction rate, the project successfully bridges the gap between technical mechanisms and ethical reflection, providing a scalable model for national AI literacy.

Introduction: Addressing the AI Literacy Gap

As Artificial Intelligence (AI) permeates critical sectors like justice, hiring, and healthcare, a dangerous gap has formed between those who develop these technologies and the citizens affected by them. Without a fundamental "acculturation" to AI, the public cannot effectively participate in the ethical debates surrounding its use.

The challenge is twofold: first, the "Black Box" problem, where AI is seen as an impenetrable mystery leading to blind trust or unfounded fear; and second, the difficulty of educational scaling. Inspired by Finland’s "Elements of AI" initiative, the IAI MOOC seeks to provide a "citizen culture" regarding AI, targeting educators, parents, and professionals who need to understand these systems to develop a critical perspective.

Pedagogical Methods and Active Learning

The MOOC employs a playful and experiential pedagogical approach rather than a purely theoretical one. Key components include hybrid delivery—combining online video capsules featuring both professional actors and experts with webinars—and active learning.

Interactive interface for testing AI image recognition capabilities.

Participants do not just watch; they "manipulate" neural networks and test image recognition programs to observe confidence levels in predictions. The curriculum covers both Symbolic AI (rule-based) and Connectionist AI (neural networks), providing historical context to show how the field has evolved. Social learning is also emphasized through forum discussions to address the "why" (ethics) alongside the "how" (technology).

A video capsule featuring a professional actor to engage the general public.

Key Insights and Philosophical Pillars

A core philosophical pillar of the course is that human responsibility is non-delegable. Delegating a decision to an algorithm is a human choice, and responsibility for the outcome remains with the human who chose to trust the mechanism.

The paper also highlights the need for citizens to understand the distinction between interpretability (understanding the internal reasoning of the algorithm) and explicability (the ability to justify a specific result to a non-expert). Furthermore, AI literacy helps citizens recognize that AI is often a specific tool for specific tasks rather than a general "human-like" mind, addressing the "moving goalpost" of what is considered intelligence.

Results and Impact

The MOOC has achieved significant scale and engagement:

  • Scale: Over 32,000 participants enrolled by early 2022.
  • Engagement: More than 10,000 active users on the forums with nearly 14,000 messages exchanged.
  • Satisfaction: An exceptionally high 94% satisfaction rate among participants.
  • Certification: Over 2,500 certificates of completion issued.

All resources are released under a CC-BY 4.0 license, allowing teachers to reintegrate the videos and activities into formal school curricula, such as the French "NSI" high school specialty.

Limitations and Future Outlook

Despite its success, attracting people who are not already interested in technology remains a significant challenge. While the MOOC is free, reaching citizens outside of educational or tech-adjacent circles is an ongoing struggle. Future work involves better-integrating follow-up resources to help learners transition from basic literacy to advanced proficiency through more technical pathways.

Summary updated

Abstract

National audience
BibTeX
@inproceedings{alexandre2023un,
  author = {Frédéric Alexandre and Marie-Hélène Comte and Aurélie Lagarrigue and Chloé Mercier and Axel Palaude and Margarida Roméro and Thierry Viéville},
  title = {{Un MOOC pour initier a l'IA : « Intelligence Artificielle avec Intelligence »}},
  booktitle = {SPIRE},
  year = {2023},
}
Figure