Abstract

The Extended Mind Thesis (EMT), proposed by Clark and Chalmers (1998), revolutionized philosophy of mind by arguing that cognition extends into the environment through tools that reliably store and retrieve information. However, the rise of generative artificial intelligence—particularly Large Language Models (LLMs)—presents a challenge that exceeds EMT's original explanatory scope. LLMs do not merely extend memory or computation; they generate novel semantic possibilities, thereby transforming the very structure of cognitive engagement. This paper argues that we have entered a new phase of cognitive evolution: the era of the Amplified Mind. Unlike the extended mind, which accesses stored information, the amplified mind navigates dynamically generated possibility-spaces. Through a reconceptualization of Clark and Chalmers' parity principle into an amplification principle, and via a new thought experiment ("Neo-Otto and Generative Inga"), I demonstrate how LLMs introduce generative uncertainty, intentional saturation, and navigational agency into the cognitive process. The result is not merely an extended mind, but an amplified cognitive architecture—one that demands new understandings of agency, responsibility, and the very boundaries of thought. Situating this analysis within the broader 4E cognition framework and Clark's own later work on predictive processing, I argue that amplification represents a genuinely novel form of mind-world coupling that existing frameworks were not designed to capture.