Abstract
Rapid advances in artificial intelligence are commonly described as technological progress, yet the growth of artificial capability and the growth of human capability are not equivalent phenomena. Artificial systems may become more capable and more deeply embedded in everyday life while the substantive capabilities, agency, opportunities, and lived conditions of the people living alongside them change weakly, unevenly, or adversely.
The paper develops a human-development conversion constraint. A public or institutional claim of AI-driven human progress is not warranted by improvements in system capability, adoption, or output alone. It requires separate evidence that technological capability has become a substantively valuable, sufficiently distributed, and sufficiently durable expansion of human possibility.
The constraint does not freeze a pre-AI list of human powers or assume that enhancement must remain inside the biological individual. It asks the claimant to identify which human powers remain necessary for meaningful agency in the altered practice, then show that the asserted gain has not been purchased by materially eroding those powers.
Preble, Micheal Charles. “The Human-Development Conversion Constraint: What Should Count as Progress in AI-Rich Societies?” Available at SSRN 7416900, 2026.