AI · SEMICONDUCTOR · FACT CHECK
SKD WIREMostly True
A classification framework dividing human capabilities into four levels in relation to AI, floated by an OpenAI founder, has drawn attention and debate. The claim — that the figure proposed a four-stage ladder for how humans stack up against AI — holds up in substance, though some details remain unconfirmed.
The framework, attributed to an OpenAI founder, sorts human capability into four tiers, framing where people stand as AI systems advance. The proposal quickly generated discussion in tech and policy circles, with critics and supporters weighing in on whether such a simplified hierarchy usefully describes the AI-era divide.
Primary sources reviewed in connection with the claim number nine, including reporting by Korea's JoongAng Ilbo alongside material referencing OpenAI-related figures Dwarkesh Patel and Andrej Karpathy. The core assertion — that the OpenAI founder publicly proposed dividing human capability into four levels — is supported by the available record.
Not every element of the framing could be fully substantiated: parts of the underlying documentation remained unreviewed due to network errors, and the exact wording and context of the proposal are not fully established.
The proposal has become a talking point in the broader conversation about AI's impact on work and skills, particularly among Korean technology audiences tracking statements from OpenAI leadership.
The claim that an OpenAI founder proposed dividing human capability into four stages in the AI era is accurate in its essentials, with minor details unresolved — making the assessment Mostly True
