The Spandrel Effect and Emergent Properties: Rethinking AI’s Non-Teleological Nature
Why does artificial intelligence feel conscious when it isn’t? The growing unease surrounding AI arises less from any machine awakening than from a deeply human attachment to teleological thinking—the impulse to see purpose, intention, and inner life wherever complexity appears. Using the Spandrel Effect from evolutionary biology, articulated by Stephen Jay Gould and Richard Lewontin, this essay reframes emergent properties in artificial intelligence as structural by-products rather than signs of agency or consciousness. By examining emergent intelligence, self-teaching systems, AI hallucinations, and contemporary debates about AI consciousness, it shows how optimization, scale, and architectural constraints generate behaviors that are repeatedly mistaken for intention. The result is a critique of popular narratives that portray AI as an autonomous or destined force, revealing instead how myths of machine consciousness function as mirrors of humanity’s enduring hunger for meaning.