Generative AI is not diminishing the value of education. It is shifting the locus of value.
For over a century, secondary schooling rewarded technical execution: writing structured essays, summarising texts, performing calculations, translating languages and applying standard procedures. Generative systems now perform many of these tasks with speed and fluency. The educational premium, therefore, moves elsewhere — towards capacities that are far more difficult to automate.
The centre of gravity is shifting from execution to judgement.
Early international signals point in three consistent directions.
1. From “doing” to “deciding”
Students must not only solve problems but also formulate tasks, evaluate outputs, define criteria and justify choices. The Brookings Institution has described this emerging demand as the cultivation of “judgment in ambiguity” (2024).
In an AI-augmented environment, knowing how to proceed is no longer enough. Students must know when, why and under what assumptions they proceed.
2. From product to process
Assessment is gradually moving towards traceability: hypotheses formulated, alternatives considered, revisions made, and reasons articulated. The OECD (2023) identifies this shift as essential for what it terms “future-proof cognitive resilience.”
The emphasis is less on the polished final product and more on the intellectual pathway that produced it.
3. Disciplinary thinking + AI (not either/or)
AI does not replace Mathematics, Biology or History. It raises the conceptual bar within them. The International Baccalaureate has already incorporated AI literacy into academic honesty and assessment guidance (IBO 2023–2025 updates), signalling that disciplinary integrity must coexist with technological fluency.
The implication is structural.
If this trajectory continues, upper secondary education may begin to resemble university sooner than expected: less reproduction, more argumentation; less execution, more criteria; fewer tasks to complete, more decisions to defend.
The schools that will thrive are not those that merely teach students to “use AI,” but those that develop students capable of thinking with AI without outsourcing their own thinking.