By Martine Rioux
Translated by Nicole Arsenault
Attempts to automate teaching are nothing new—and have mostly fallen short (Rensfeldt & Rahm, 2023). Yet with the rapid rise of generative AI tools like ChatGPT, new speculation has emerged: Could these tools transform—or even replace—some educational functions?
So far, few studies have tested these technologies directly against the real demands of the teaching profession. That’s exactly what Alexandre Lepage, Samira Karim, and Florent Michelot from Concordia University set out to do. They presented their findings (in French) at the recent International Education Conference hosted by the Interuniversity Research Centre on Teacher Education and the Teaching Profession (CRIFPE) in Montreal.
To explore this question, the researchers relied on the Reference Framework for Professional Competencies for Teachers (in Québec), an official document that guides teacher training and professional development.
For each competency, they identified key terms, conducted a review of academic literature, and analyzed documented uses of AI in educational contexts. These elements were then reviewed by the three researchers. Each independently assessed the automation potential of every competency—low, average, or high—before comparing their evaluations to reach a consensus.
The verdict? Of the 13 competencies assessed, only four were deemed to have high automation potential:
Among these, evaluation (or assessment) is the most discussed in academic literature, particularly given recent advances in generative AI capable of producing feedback or analyzing performance. But even in these cases, the researchers emphasize: this is an analytical perspective—not an invitation to blindly automate these tasks.
The remaining competencies—especially those related to classroom management, student support, and collaboration with families—are likely to resist automation much longer. The researchers explain that this is due to the complexity of the profession, which relies on relational, contextual, and reflective skills that are difficult to model.
Interestingly, the researchers noted that their study also highlighted the limitations of the professional competency framework itself. “We realized that the vocabulary and descriptions are often too vague. In short, the framework poorly reflects the reality in the classroom. There’s a disconnect between ministry documents and day-to-day practice. So, it’s difficult to make a detailed evaluation and claim we’ve truly analyzed teachers’ daily work.” (translated from quote in French). They believe this insight could contribute to a wider discussion about what it means to be a teacher in the age of AI.
The researchers suggest several avenues for further research: validating their tool with a sample of teachers, exploring the concept of agency in the human-machine relationship, and better understanding the systemic—and unexpected—impacts of using AI in the classroom.
In short, AI doesn’t seem ready to “pass its student-teaching internship,” a nod to the formal teaching practicum process. White it can support some aspects of the profession, it cannot replace the human depth and richness of teaching itself.
If you are subscribed, log in to access exclusive member content!