AI threatens cultural monopoly, sparking control debate

AI threatens cultural monopoly: who controls the narrative?

The promise that anyone can generate a movie, series, or video game from home using a powerful computer is no longer science fiction. In recent years, generators for images, music, and text have evolved from expensive toys into accessible tools. Some economic analyses suggest the next leap—full audiovisual production—is two or three decades away, perhaps less. The thesis driving this debate is uncomfortable for the cultural establishment: if content production costs drop to near zero, the narrative monopoly held by a few platforms and studios cracks. This is not a neutral prediction. It is a political bet disguised as technological analysis, and it should be read with that warning in mind.

Production cost as a lever of power

The central argument is simple: whoever controls the means of production controls the message. For decades, making a film or series required million-dollar budgets, large teams, and global distribution. According to this view, that bottleneck has functioned as an ideological filter. AI promises to break it. With a high-end PC and open-source models, an individual creator can generate textures, synthesized voices, maps, scripts, and basic animations. The result won't be The Godfather, but neither is 90% of what is produced today with public and private funding.

Some argue the true revolution isn't quality, but the quantity of voices. If the marginal cost of producing a series approaches zero, the number of creators multiplies exponentially. Along with them come perspectives that currently find no space on dominant platforms. The most optimistic calculation assumes this diversification erodes the single narrative; the pessimistic scenario starts from the premise that the generated noise will be so massive that attention will remain concentrated in a few trusted nodes.

Model censorship and the corporate counterattack

The first serious objection to technological optimism is that current models are restricted. Large language and image generation systems filter, bias, and block content according to opaque editorial criteria. A user can verify how a request about an uncomfortable historical topic receives a refusal, while another about a twentieth-century dictator is resolved without issue. That asymmetry is not an accident: it is design.

The counterargument is that technology cannot be contained. Open models already circulate, and each version consumes fewer resources than the previous one. What required a data center two years ago now runs on a home tower. The question is whether this race between openness and control will tip toward one side or the other, or if two ecosystems will coexist: one corporate, sterile, and monitored; the other decentralized, chaotic, and difficult to monetize.

The human limit: talent, not hardware

The least discussed factor is the human element. Having access to a tool does not equate to knowing how to use it. Most AI-generated products will be mediocre, just as most self-published books or homemade videos are. The difference is that a talented minority will face fewer barriers than ever before. The comparison with the printing press is recurrent: it multiplied books but did not turn all readers into authors. What it did do was break the scribes' monopoly.

This is where the analysis gets stuck. If talent remains scarce and attention remains finite, AI's real impact on the cultural balance may be smaller than euphoria suggests. Or it may be greater, but through indirect paths: not because geniuses emerge, but because gatekeepers disappear.

---

The discussion is not closed. Those who see AI as a tool for cultural liberation clash with those who recall that the internet promised the same thing and ended up concentrated in four platforms. The difference, the former argue, is that hardware and models are now in everyone's hands. The difference, the latter reply, is that power always finds a way to charge a toll. With these elements, the next decade will determine whether the basket is woven by an open algorithm or a monthly subscription.