Token Inequality Revives Korea's Knowledge Divide

The article argues that the cost structure of AI, where users pay per token for computing and reasoning, is reintroducing a form of knowledge inequality. Unlike the internet era, where flat-rate access helped narrow the information gap, AI's usage-based fees mean that those with greater purchasing power can engage in more extensive analysis, refinement, and iteration, potentially leading to superior outcomes in education and business.

This token divide has direct implications for fairness in competitive environments. For students, the ability to use AI for multiple drafts or to compare different models can significantly impact assignment quality. Similarly, startups developing prototypes or verifying code may find their ability to experiment and refine limited by token budgets, making it difficult to distinguish between technical capability and financial constraints when evaluating project outcomes.

The core distinction highlighted is that while the internet lowered the barrier to accessing information, AI broadens the capacity to analyze and apply that information as knowledge. However, this enhanced capability is now tied to a variable cost, meaning that the depth and breadth of AI-assisted learning and development are increasingly contingent on one's ability to afford continuous usage.