AI could launch us back into the middle ages, and that's a good thing.

Summary

The speakers discuss how AI may threaten human agency precisely because it presents itself as helpful—offering convenience and power rather than appearing as an obvious danger. Jonathan Pageau argues that concern about the future can be useful if it motivates people to prepare and act, rather than despair. Meeting this challenge requires discipline and resisting the desire to outsource everything.

They connect this concern to education, which Pageau says should preserve continuity between generations. He describes three guiding principles: remembrance, celebration, and creation. Education should connect subjects rather than present isolated facts, approach history and texts sincerely rather than cynically, and equip students to carry knowledge forward to future generations.

Because AI can generate student work, Pageau supports returning to methods that test what students actually know, such as oral examinations and handwritten, in-class work. He warns that if teachers and students all rely on AI for lessons, assignments, grading, and assessment, education risks becoming a closed loop in which people receive grades without meaningfully teaching or learning.

Key Strategies and Recommendations

• Treat pessimism as a prompt to prepare and make changes—not as a reason to despair.
• Stay aware of risks while avoiding needless worry.
• Practice discipline by resisting the temptation to let convenient tools do everything.
• Teach knowledge in context, showing how subjects connect and fit into a broader human story.
• Approach what is being taught with sincerity rather than default cynicism.
• Teach with the aim of preparing students to contribute and pass knowledge on to the next generation.
• Use oral examinations and handwritten, in-class work to assess students’ understanding directly.
• Value what students have learned and can explain themselves, alongside their ability to research and find references.

Presenters/sources: Jonathan Pageau; interviewer not identified.