When AI Knows Everything, What Should Humans Learn? | Kristina Kallas | TEDxUniversity of Tartu

Overview

Kristina Kallas argues that even though AI can perform many knowledge-based tasks, the real educational challenge is how to evolve human learning and thinking skills in an era of cognitive pressure and “cognitive offloading.”

Estonia’s policy approach

Kallas explains that Estonia’s education ministry began addressing AI policy only after students started using it in schools, following the spread of ChatGPT in late 2022. After months of consultations with experts (including neuroscientists and cognitive scientists) and industry, her team concludes that education policy should not focus on the technology itself—but on what AI means for human learning capacity, especially the brain’s cognitive development.

Key educational concern: higher- vs. lower-order thinking

A central theme is that current education systems historically emphasize:

• Lower-order thinking: memorizing, understanding, applying
• Less emphasis on higher-order thinking: analysis, evaluation, ethical judgment, and creation

Kallas claims that in an AI era, relying heavily on lower-order skills becomes insufficient because AI can already handle much of remembering and applying knowledge. She further warns that without deliberate instructional change, schools may drift toward inertia—leading to “brain rot” or excessive offloading.

Why change is needed (historical comparison)

To justify this shift, Kallas compares today’s AI pressure to past technological revolutions—especially the invention of the printing press, which eventually led humans to learn reading universally.

Similarly, she argues AI creates a new evolutionary cognitive pressure: humans must become faster, more systematic, and more creative thinkers.

The brain’s modes: lower vs. higher cognitive mode

Kallas describes the brain as operating in two modes:

• Lower cognitive mode: energy-saving, automated, repetitive learning
• Higher cognitive mode: energy-intensive, requiring relearning and complex thinking

Since AI makes constant low-mode functioning less sustainable, education must train people to spend more time using higher-order thinking skills—starting in lower secondary school.

Estonia’s response: the “AI leap” program

Kallas outlines Estonia’s response: an “AI leap” program that integrates AI as a tutor and assistant designed to push students toward:

• analytical thinking
• systematic thinking
• creative thinking
• critical thinking

How it is structured and funded

• Implemented as a private–public partnership
  • Half funded by private companies
  • Half funded by the ministry
• Run by a foundation, rather than the ministry
• Starts with upper-secondary students (grades 10–11) in a pilot phase
• Planned to extend to vocational education students

Tutoring tool and monitoring

Kallas emphasizes that Estonia uses a Socratic-style AI tutoring tool developed with OpenAI—not the generic ChatGPT, but a school-focused Socratic version.

The program includes scientific monitoring, though she notes that research on AI’s effects on learning is still limited. Estonia is therefore observing and learning as the initiative develops.

Core goal: “wise” use of AI

The goal is not simply to increase time spent using AI, but to use it wisely—to guide learning toward higher-order cognitive skills and to generate scientific knowledge that others can study.

Presenters or contributors

• Kristina Kallas (speaker, TEDxUniversity of Tartu)
• Laila Agha (subtitles transcriber)
• Asqat Yerkimbay (subtitles reviewer)
• OpenAI (partner mentioned for R&D of the Socratic tutoring tool; not a specific individual)