What is the difference between the technological and the human? It seems as though we should be able to draw up a balance sheet with two clear columns: typewriters on one side, humans on the other. However, radical postmodern thinker Katherine Hayles argues that our notion of humanity and its separation from what is artificial, synthetic or ‘computational’, is an illusion. In this article, Hayles explains how the ‘artificial’ and ‘natural’ columns are merged into a single inter-braided mixture.
In this era of advanced artificial intelligence, perhaps the most highly charged binary is between an AI (on the technological side) and a human (on the biological). Leaving aside the trend to incorporate more and more technology into human bodies, including pacemakers, insulin pumps, cortical and neural implants, cochlear devices and so forth, humans and our cultures are forming strong symbiotic bonds with computational media in all their forms. There is a reason for this: computational media are the first technologies to have cognitive abilities. Cognition, understood as the ability to sense information from the environment, interpret it, and act in contexts that connect it with meaning, emerged first on Earth in biological bodies. Even one-celled organisms such as bacteria move toward food sources and away from toxins in ways that enable them to continue their existence. In fact all lifeforms, even those that don’t have brains such as plants and clams, have cognitive capacities. Now humans have found ways to create cognition in artificial media, leading to complex interdependencies between cognitive media and cognitive humans in collectivities that I call cognitive assemblages.
Biologist Lynn Margulis has argued that symbiosis is the primary driver of biological evolution, based on her then-startling (now accepted) theory that mitochondria and chloroplasts are remnants of previously independently living organisms that had been absorbed by cells, only partially digested, and eventually merged so they became part of the cells proper. Margulis has called this the most important event in Earth evolutionary history, because it allowed eukaryotic cells to use aerobic respiration, providing greater energy efficiency and paving the way for the emergence of plants and animals. Through a long process of evolving increasingly more complex organisms, evolutionary dynamics led to the emergence of Homo sapiens, the cognitive species par excellence, capable of building what is now an extremely deep technological infrastructure.
Sara Walker, an astrobiologist and theoretical physicist, has used this lineage to argue that “AI is life,” because it emerged from life.
How did this infrastructure evolve? Anthropologists define a tool as an artifact used to make another artifact. Imagine how difficult it was to create the first tool (probably a stone flake). No doubt it was maddingly slow work, because by definition, there was no other tool to assist in its manufacture. Once one has a sharpened rock, one can make a stone axe; having a stone axe makes it possible to hollow out a canoe; having a canoe increases mobility and widens the horizon of gathering and hunting, and so on. Each successive invention builds on prior ones, at exponentially accelerating rates. By the third millennium CE, humans have discovered ways to offload memory, information, and cognition into digital computers that can store them, transform information into different codes, manipulate the codes, and transmit the results worldwide in a few seconds.
John Maynard Smith and Eörs Szathmáry have argued that major transitions in biological evolutionary trajectories on Earth are marked by profound transformations in information storage and transmission. Echoing Margulis’s theory, they also argue that a consistent pattern in such transitions is that some entities lose the ability to replicate independently, reproducing themselves only as part of larger wholes. Although humans can still reproduce on an individual scale (thank goodness!), human societies have become so dependent on advanced technologies that they cannot replicate themselves at scale, for example with terraforming, without the deep technical infrastructures controlled and guided by computational media. Transportation networks, food supply lines, communication and electric grids, and thousands of other necessities are now completely dependent on such cognitive technologies. Human cultures are on the cusp of a new era in entwined biological and technology evolution, in which the symbiosis between computational media and humans morphs into technosymbiosis, a process that produces the cognitive assemblages that do most of the developed world’s work, in which humans, nonhumans and cognitive media cooperating to accomplish increasingly complex tasks.
This new technosymbiosis has at its heart a technology that does something no other (nonhuman) species or (human) technology has been able to do before: create complex linguistic texts that are semantically and syntactically appropriate in response to human inputs. These generative AIs, known as Large Language Models (LLMs), use neural net architectures inspired by biological neurons and operating at very large scales. The potentials of these technologies for human welfare are amazing; already they have been used to predict protein foldings that can be used to develop life-saving drugs. Of course, they also have potential to do real harm if they are appropriated for malign purposes by bad actors. There is also the possibility that as they learn, they may develop a technological equivalent to awareness and begin to implement their own purposes rather than those of their human masters—Terminators for real.
Human cultures are on the cusp of a new era in entwined biological and technology evolution, in which the symbiosis between computational media and humans morphs into technosymbiosis.
Even as we vault into this unknown territory, it is important to realize that we have not left our evolutionary past behind. The emergence of eukaryotic cells is instantiated and preserved within our bodies; also present are all the later evolutions that enabled multicellular organisms to emerge from unicellular ones, including the processes of cell differentiation into organs and tissues, the hallmark of more cognitively complex organisms. At the apex is the human brain, with its 86 billion neurons and 100 trillion synaptic connections, which has enabled perhaps the most important evolutionary event since eukaryotic cells: human language and symbolic thought.
Now that emergence has led to another spectacular evolutionary transition, the AI devices of LLMs that are trained in billions of human-authored texts and can use language in meaningful ways themselves. In a certain sense, one can say that there is a lineage here that goes all the way back to eukaryotic cells, since these advanced AIs would not exist if it were not for human brains, which would not exist were it not for multicellularity, which would not exist were it not for eukaryotic cells with mitochondria whose metabolic processes power cells/tissues/organs/brains/humans, which in turn supply the electrical currents that power AIs.
Sara Walker, an astrobiologist and theoretical physicist, has used this lineage to argue that “AI is life,” because it emerged from life. Thus her argument deconstructs the biological/technological binary completely by following what she calls the “information lineage” of advanced AIs. Sympathetic to this approach, I recognize that there are some large leaps that would need to be closely argued to be convincing. Walker traces a similar information lineage from unicellularity to multicellularity to human societies and then to multisocietal collectives, suggesting that we are on the verge of another momentous transition from the individual to regional clan to the nation-state to the planetary. The vision is enticing, but much will depend on whether human-technical cognitive assemblages can save our planet by summoning the political will to institute controls and institutions at the planetary scale. Given our responses so far to our multiple anthropogenic environmental crises, optimism may not be warranted. Peering into the cloudy glass of our futures, I am confident of only one thing: barring environmental collapse and assuming continuing technosymbiosis, the evolutionary trajectories of humans and AIs are completely entwined.