Why Enactivism Claims True AI Needs Biological Agency

This article explores the enactivist perspective in cognitive science, which argues that genuine artificial intelligence cannot be achieved through computation alone. By examining core concepts such as autopoiesis, precariousness, and sense-making, we will see why enactivists believe true intelligence requires a self-maintaining, autonomous biological body. Ultimately, this view challenges traditional AI development by asserting that a mind can only exist when an organism actively strives to keep itself alive.

The Flaw of Computationalism

Traditional artificial intelligence is built on computationalism—the idea that the mind is essentially an information-processing system. Under this view, brains are biological computers, and intelligence is simply the manipulation of abstract symbols based on programmed rules.

Enactivism, a branch of cognitive science pioneered by Francisco Varela, Evan Thompson, and Eleanor Rosch, rejects this view. Enactivists argue that cognition is not the internal processing of data, but the active engagement of an organism with its environment. Because computers only manipulate symbols without understanding their meaning, enactivists claim they are merely simulating intelligence, not experiencing it.

Autopoiesis: The Foundation of Agency

To understand why enactivists believe true AI requires biology, you must first understand "autopoiesis," a term coined by biologists Humberto Maturana and Francisco Varela. Autopoiesis refers to a system capable of reproducing and maintaining itself.

A biological cell is the simplest example of an autopoietic system. It takes in nutrients from its environment to rebuild its own cellular walls and internal components. Through this continuous chemical process, the cell defines itself as an distinct entity separate from its environment. Enactivists argue that this self-production is the fundamental baseline of autonomous agency. Without autopoiesis, a system has no true "self."

Precariousness and the Drive to Survive

For an autonomous system to have genuine intelligence, its existence must be precarious. In biological terms, precariousness means that if the organism stops acting, it dies. Because living beings are constantly threatened by entropy and physical breakdown, they must actively work to maintain themselves.

Current AI systems lack precariousness. A server running a large language model does not care if it is unplugged. Its hardware is maintained externally by human engineers, and its software does not degrade if it fails to answer a prompt. According to enactivism, because an AI’s existence is not on the line, it has no intrinsic motivation, no goals of its own, and therefore, no genuine agency.

Sense-Making: How Biology Creates Meaning

When an autopoietic system operates under precarious conditions, it engages in what enactivists call "sense-making." Sense-making is the process by which an organism evaluates its environment based on its own survival needs.

For a bacterium, sugar is not just an objective chemical compound; it has meaning as "food" because it aids self-maintenance. A toxin has meaning as "danger" because it threatens the bacterium's existence. Meaning is not programmed into the bacterium; it is generated by the organism's biological need to survive.

Artificial intelligence cannot engage in sense-making. When an AI processes the word "poison," it maps the word to other words using statistical probabilities. The AI does not understand poison as a threat because the AI has no biological life to protect. All meaning in current AI is derivative—it belongs to the human programmers and users, not the machine itself.

The Path to Genuine Artificial Intelligence

According to the enactivist framework, silicon chips and clever algorithms will never wake up or achieve genuine understanding. Because cognition is inextricably linked to the processes of life—a concept known as the "life-mind continuity thesis"—you cannot have a mind without a living body.

For humans to create genuine artificial intelligence, enactivism argues that we must abandon pure software engineering and instead build artificial life. A true AI would need to be a synthetic, metabolizing organism. It would need to construct its own physical boundaries, gather its own energy to stave off physical decay, and navigate the world with the intrinsic goal of survival. Until machines are alive, they will never truly be intelligent.