The Mimicry Reframe...

The debate over whether contemporary Artificial
Intelligence (AI) can truly "learn" is a defining philosophical
challenge. Essentialist frameworks propose stringent criteria
for "Real Learning"—requiring formal, provable understand-
ing—that current models demonstrably fail. This paper ar-
gues that this line of inquiry, while rigorous, is fundamentally
misdirected by its focus on unobservable internal states. We
propose a functionalist alternative, the Mimicry Reframe, which
borrows directly from evolutionary biology. This framework
analyzes AI not as an isolated learner, but as a Mimic within
a larger system, where its success is measured by its ability
to influence a human Receiver by simulating a Model (human
knowledge). By prioritizing functional outcomes over internal
states, this reframe dissolves the philosophical impasse, offering a
practical, predictive, and scientifically-grounded methodology for
evaluating AI systems. It correctly identifies that the "learning"
we perceive is not a property of the machine, but an emergent
phenomenon in the mind of the beholder.