How AI Shapes Us: McLuhan And The Grammar Of LLMs

The prophet McLuhan

Every era chooses its own determinations. At one time, it was thought that there was a fate assigned by the gods, or by a single God. In other moments, it was the laws of history, will, reason, or race, the subsystems that condition the economy or class struggle. More recently, the determination was linguistic, cultural, or gender-based. In the present, although elements of all these beliefs persist, the first determination we think of is technological. Hence the discussions about the power of algorithms, about our inability to be happy, which stem from our use of the like button, or the complaints about the destruction of democracy due to information niches, the echo chambers produced by social networks, and so on.

In this obsessively human quest to identify our new determinations, the work of a 20th-century thinker who had found himself in a rather uncomfortable position gains new value. The Canadian Marshall McLuhan (1911-1980) was the first to identify and articulate the new era of media that had solidified with the popularization of television in the mid-century. He became extremely famous. He appears in Annie Hall debating with Woody Allen about one of his theories, and in the excellent and necessary Videodrome (1983) by David Cronenberg, the media guru “Brian O’Blivion” is directly inspired by his figure and ideas. But this popularity also came at a cost. Firstly, his essayistic and somewhat unacademic style (and at times rather repetitive), along with grandiose and radical statements, made him less digestible for academic audiences. Secondly, the type of generalization he was very fond of left many openings for criticism, making it easy to point out his inaccuracies and argumentative leaps. This was done, among others, by Raymond Williams, the father of British cultural studies. Thirdly, McLuhan did not fit well into the intellectual climate of the 1960s and 70s. He was not a Marxist, he did not focus on denouncing the powers that be, he self-identified as Catholic, and he described the changes brought about by media with a tone of inevitability that often seemed celebratory. The effects that, according to him, technologies have on human consciousness were not something that could be judged as “negative” from an ivory tower. He was far from the fierce critique of the “culture industry” (understood as the ideological manipulation of the masses) carried out by Theodor Adorno and his direct and indirect disciples.

The fundamental point of reference for beginning to understand McLuhan is what he is most famous for, the phrase “the medium is the message.” Far from meaning something like “the ideological line of a media company is expressed in all its content,” which is the sense in which it is most commonly used, the power of this idea lies in the fact that for McLuhan, media are “extensions of the human being.” Just as television extends (as its name literally indicates) sight, the wheel extends the foot. But is the wheel a “medium”? We usually use that word to refer to mass communication media, whether as concrete technologies (radio, television, internet) or companies (Clarín, La Nación, Gelatina), but one of McLuhan's great theoretical bets was to broaden its meaning to include any “device” (in the broad sense that the phonetic alphabet is a device) that allows us to expand our senses and capabilities as a species.

So, what does it mean that the medium is the message? It means that what matters about a medium is not what it says, but what it does, and that dimension of doing can never be reduced or based on content. Technologies affect us at a much deeper level than our political opinions or musical tastes. They change our sensory relationship with the environment, with others, and with ourselves, both in terms of individuals and culture or species. Human beings are what media make of them. The medium is never neutral. A gun can be used to shoot an innocent man or an enemy army, but it is always used to shoot, expanding the capacity to harm of the human body. A social network can be used to meet friends and get reading recommendations or to radicalize and become a Nazi, but it is always operating in the same way on your mind and body. Broadcasting educational programs on television does not make it any less television. The content of a medium is like a sausage thrown to the guard dog so it doesn’t see that someone is breaking in, McLuhan dixit.

We will not attempt here a summary of McLuhan's central ideas. To discuss how his work helps us think about present artificial intelligence, it is more productive to identify three ideas found in his work (some more developed than others) that have a significant impact on contemporary media theory and theories about the effects of technology.

First, as we have already anticipated, media are not passive transmitters of messages or pre-existing content. Analyzing a medium involves being able to see beyond what it shows on its surface, but not to look for the “hidden message” (as ideological critique typically does), but to understand how it affects our sensory and psychic apparatus. For McLuhan, media not only amplify. They also numb or even “amputate” faculties and capabilities. The increase of one aspect always results in the detriment of another. Contemporary hypersensitivity is also an anesthetization.

Second, we should not understand technologies and media as mere consequences of progressive technical projects and developments. That is, a medium is not exhausted by the history of the previous inventions that made it possible, nor is it explained by the declared intentions of its creators. The conditions of possibility are always more complex and often form part of systemic problems of a culture or civilization, many of which (if not all) only become visible long after their emergence. Moreover, media do not only make something pre-existing obsolete: they also recover and incorporate something from the past.

Third, media shape the world, but they do so in a quasi-invisible manner. They are part of the permanent backdrop of our interactions, our self-perception, and our way of making sense of what surrounds us. They are like water, and we are the fish, which is why we do not see them. Of course, we see media as such (like the PC I am writing this on), but what is in the foreground (the fact that I can see the pixels on my monitor reproducing the letters I type on my keyboard) never fully represents the effects they produce. Immersed as we are in an electrified, digitized world covered in screens of various sizes, only through effort can we begin to understand what technique has made of us.

The algorithmic extension of language

The popularization of the Internet has already brought about a revitalization of McLuhan, particularly due to his famous metaphor of the “global village” (a world community that, by being hyper-connected, recovers elements of pre-modern tribal societies), which is much more suitable for the web than for the black-and-white television world in which it was formulated. Today, with the massification of chatGPT and its competitors like Gemini, Claude, DeepSeek, and a couple more, we must ask ourselves what we can understand about generative artificial intelligences (GAI) if we look at them through this framework.

In principle, saying that GAI is a “medium” encounters the problem that arose with the advent of the personal computer (connected to the internet, as all computers are today) to traditional communication theory, which was accustomed to separating its effects based on concrete technologies like printing, radio, cinema, or television. The computer is a “metamedium” in which we do, among many other things, what we did with previous media, such as listening to the radio, watching a movie, or sending messages. The GAI that 99% of humanity uses is nothing more than an application, a program that runs online and has a particular interface that connects with the technology that operates behind it, the famous “large language models” (or LLMs) in all their layers of software and hardware.

If for McLuhan the alphabetic writing was an extension of the voice and the wheel an extension of the foot, generative artificial intelligence could represent the extension of the very process of cognition and language. We are no longer just externalizing memory (as we did with writing and its various supports) or the speed of calculation (with traditional computing), but the ability to synthesize, associate, and articulate meaning. In doing so, we may be anesthetizing the epistemic capacity of writing. Thought is not something that occurs “inside” and then is “poured” into words; there is a significant part of it that happens in the very act of writing or speaking. The “message” of GAI is not the content that the model responds with, but the reconfiguration of our relationship with truth and authorship. In a world where content is generated by a “stochastic parrot,” the notion of style becomes an adjustable parameter, just another filter. The GAI medium is telling us that language is a perfectly predictable resource and, therefore, automatable. We are witnessing the transformation of culture into a database that, unlike books placed in a library (which depend on human readers capable of putting them in dialogue), can generate meaning by itself.

McLuhan liked to invoke the myth of Narcissus to explain that the young man did not fall in love with himself, but with an extension of himself that he did not recognize as such (his reflection). When interacting with LLMs, we experience algorithmic narcissism that, unlike social networks, has the capacity to completely isolate us from the gaze of the other. When we read a response from the AI, we are seeing a processed reflection of the human collective consciousness, but it is a consciousness that presents itself as eternally friendly and available to us, the opposite of that “Big Brother” that terrifies us in total control dystopias. The new global village is a labyrinth of mirrors. The medium, instead of projecting an external world (however manipulated it might be), projects a version of reality hyper-adjusted to our semantic and pragmatic expectations.

Understanding GAI through McLuhan forces us to stop looking at the screen and look at the environment. The true effect of AI is not in the quality of the content it produces; it is in how it changes the structure of our work, our education, and our perception of what it means to be human. If human beings are what media make of them, GAI is shaping a subject that seeks synthesis and personalization rather than depth or the revelation of something that transcends it. But that does not mean we should fall into pessimism. In the world of media and technologies, there is no amputation without the expansion of another capacity. It is possible to think that we are facing a radical extension of our orchestration capacity. If the man of the “Gutenberg era” (dominated by printing, linear cause-effect relationships, and individualism) was a specialist dedicated to deepening a single line of thought, the subject of the AI era resembles more a conductor. Intelligence ceases to be an accumulation of data or a mechanical writing skill to become a faculty of design, of knowing how to ask questions, and of knowing how to connect results that were previously scattered in the background noise of global culture. The content of GAI is precisely all the previous media culture processed and made available as a set of probable associations. What we are experiencing is not the end of human creativity. Rather, it is its transition from a craft phase to a systemic phase. Just as the calculator did not make us worse at math, GAI could free us from the “carpentry” of language to allow us to inhabit the architecture of ideas.

Technology is not something we simply use; it is the environment in which we exist and with which we co-evolve. In this new evolutionary leap, the challenge is not to resist automation, a gesture as futile as that of someone trying to continue sending messengers on horseback in the age of the telegraph. It is rather to understand the new grammar of existence that these models generate both at the surface and background levels. Freedom does not come from taking a critical distance from the outside, but from accepting that we can only understand technical flows by allowing them to transform us, honing the sensitivity necessary to inhabit their speeds and turbulences without becoming adrift.