HAVANA TIMES – Thinking back, writing allowed us to hear an absent voice. Photography brought us close to distant objects; cinema added movement to the images, and television put us in close contact with events happening thousands of miles away. The internet further reduced distances and multiplied the possibilities of influencing public reality.
In his 1927 film “Metropolis,” Fritz Lang imagined a future that could be interpreted as a reflection of the present. Lang didn’t need to predict the world that was coming – it was enough to portray the ultimate consequences of the already existing power structures. Fiction doesn’t necessarily need to anticipate reality; it can create a space through which reality begins to become visible. Art creates cracks where the light filters through.
“Metropolis” can be seen as a Platonic allegory: the workers live in an underground ghetto at the heart of the industrial complex and are forbidden to venture out to the world beyond. Incited by a robot, they rebel against the intellectual class that holds power, threatening to destroy the city up above. Freder (Gustav Fröhlich), son of the Metropolis leader, and Maria (Brigitte Helm), of humble origin, try to prevent the destruction by appealing to love and emotion.
The advent of the camera radically altered our relationship with representation. Although the term “realism” has much older philosophical roots, it was in the nineteenth century—coinciding with the era of photography—that it began to take shape as an aesthetic and artistic category with a new aspiration: to approach reality through its representation. The camera seemed to offer something that painting could not guarantee in the same way: an image produced from the physical presence of what had been placed in front of the objective. That apparent fidelity also transformed our perception; now we weren’t just representing reality but beginning to view it through the machines that captured it.
A large corporation engaged in developing artificial intelligence has buildings, laboratories, servers, shareholders, employees, and executives. It has a power structure. But whereas power has historically required a visual representative, it can now find in AI a way to conceal its own body. The machine appears to make decisions, although behind it are those who determine its limits, its objectives, and the conditions under which it operates.
In *Metropolis*, the machine has a heart. In our time, perhaps that heart is distributed among data centers, algorithms, communication networks, and the human bodies that make its existence possible. The difference is that it is no longer so easy to identify where the machine ends and the human being begins.
Lang’s robot is a stand-in. But that same threat is back in today’s world—one that may not consist of machines replacing humans, but rather of humans beginning to organize themselves like machines: obeying systems they do not understand, accepting decisions they cannot question, and delegating their own perception to an intelligence that reflects back to us an image of ourselves.
A hundred years later, the fundamental question may not be who will dominate the future, but who will have the power to define what we mean by reality—because it is no longer just a matter of representing reality, but also of augmenting it, assuming it, and modifying it. The expression “augmented reality” then acquires a meaning that goes beyond technology: what happens when what we perceive as reality contains ever more elements produced by machines? And who decides which of the added layers we should except as true?
The appearance of algorithms has brought into play another form of presence in which identity can remain operative even when the body is absent. Digital traces made it possible to reconstruct an image of a person, link it to others, and circulate it. An algorithm is, in a sense, a decision turned into a procedure, mathematics made concrete. It not only organizes information but also plays a role in selecting what we see.
The internet seemed to promise a voice for everyone. But algorithms introduced another condition: not everyone would be seen. Visibility depends on multiple factors: platforms, networks, resources, the location from which one speaks, and the institutions capable of legitimizing a discourse. It is not the same to post from New York, Madrid, or Havana. The ability to disseminate your post, access certain media, or establish international connections also plays a role in the construction of the public figure. Inequality doesn’t disappear when everyone can post; it merely changes form.
Why, after all this time, do I feel the need to understand something I had access to but previously felt removed from? The fact is that artificial intelligence adds a new layer, and that layer has caused me to distance myself. AI doesn’t merely distribute a person’s digital footprints: it can interpret them and return them organized into a coherent representation. When asked who someone is, the machine produces an answer that seems to bring together that person’s essential features. But – does what you can find really reflect what’s essential?
The intergovernmental Organization for Economic Cooperation and Development has warned about inequalities in access to data, computing power, talent, and the infrastructure needed to develop artificial intelligence. These inequalities also raise questions about who is represented in the data and who remains underrepresented.
The social groups and cultures best represented in artificial intelligence systems tend to be those with the greatest presence in the original training data. English occupies a dominant position, and content produced in other languages is unequally represented. There are also biases in the representation of social groups: some models, for example, more frequently associate men with prestigious professions and women with domestic tasks.
Many indigenous languages and other languages with few digital resources are underrepresented. Furthermore, this uneven presence does not guarantee an accurate representation: a culture may appear in the data and still be reduced to stereotypes. The machine perceives the world through a selection of the digital traces that societies have produced and preserved through this medium.
I asked Chat GPT-6 what kind of person it had constructed based on the digital traces it found of Alexei Navalny. The response brought together a series of public identities: Russian opposition figure, anti-corruption activist, blogger, researcher, candidate, political prisoner, Putin critic and founder of an anti-corruption organization. The decisive events of his life were also cited: his poisoning in 2020, his return to Russia in 2021, and his imprisonment and subsequent death in prison in February 2024.
The machine had organized a life based on its most recognized moments, and I began to select these fragments in order to construct my own interpretation. AI wasn’t the only agent doing this – I did it too.
Navalny did not appear in isolation, but rather within the political machinery surrounding him: the Russian state, Putin, the opposition, and persecution. From within that structure, Navalny appeared as the man who stood up to Power. It even began to seem to me like a macabre script in which Navalny played Navalny.
Then I also wondered about the relationship between Navalny and the large digital audience watching him. His statements circulated, his messages received responses, and his challenge to power found an international audience. To what extent did that attention help solidify a public image from which it was no longer easy to escape? What happens when the online image of a person begins to take on a life of its own, independent of the person who embodies it?
Left out was the human being, that which could not be reduced to his function within that machinery. What was the difference between the digital Navalny and a machine, when both are represented based on the place they occupy in a system? Where were the moments without public significance, the small experiences that did not need to become events to be part of a life? Perhaps we humans have long been living within representations that shape our identity and determine how we are recognized. When that image takes on the appearance of objective knowledge, we may forget that it is also an interpretation.
Artists, intellectuals and politicians have long been aware of the gap between the person and their public image. They were able to attempt to maintain the boundary between these, even if they didn’t always succeed. Now, though, an adolescent can come up against an image of themselves on the internet that is the object of others’ interpretations, accumulates comments, and remains accessible even when they themselves have changed. Are we prepared as a society to coexist with a public alter-ego that escapes our reach?
We have the tools to produce and disseminate images, but that doesn’t mean we’ve learned to live with the consequences. Public identity can grow faster than our ability to understand it, question it, or detach ourselves from it. “There is no act more cruel than looking through the lens of a camera.” Perhaps, in this digital age, the oversaturation of images is a new mask to protect our virtual selves. Or, rather, to bury what we don’t like, even though algorithms and AI may preserve its traces.