Is Learning Stealing? Tabula Rasa, Human Creativity, and Generative AI

Abstract

Creativity has historically emerged through interaction with inherited knowledge, cultural traditions, artistic conventions, language, technologies, and prior creative works rather than from an intellectual tabula rasa. Yet the rapid expansion of generative artificial intelligence has intensified debates over whether computational learning constitutes legitimate learning, imitation, or unauthorized appropriation. This article examines these concerns by challenging conceptions of human originality that treat creativity as creation independent of preceding influences. It compares human learning and generative AI as processes involving exposure, pattern recognition, recombination, adaptation, and transformation while emphasizing critical differences in embodiment, intentionality, agency, scale, speed, memory, and technological mediation. The analysis then considers how these differences complicate conventional understandings of authorship and intellectual ownership when models are trained on extensive collections of cultural and creative material. Rather than equating human cognition with machine learning, the article reframes the debate around the conditions under which learning becomes ethically or legally problematic. It argues that meaningful evaluation of generative AI should therefore prioritize transformation, attribution, consent, provenance, ownership, social legitimacy, and originality when determining boundaries between learning, creative influence, derivation, and appropriation.