I’ve noticed that the website where I’ve been publishing my articles for the past 27 years has added a tool to indicate the percentage of a published text that was generated by AI. For decades—on this site and in dozens of newspapers from different countries, a lifetime before commercial AI—I’ve published nearly 900 articles, averaging 1,200 words each. Texts—like those of my fellow writers and friends—that will become the medieval incunabula, Egyptian papyri, and Sumerian clay tablets of the future. In other words, they are unlikely to lose value over time; quite the opposite: they will be the texts of the era when AI had not yet shaped human thought and emotions, with all their traumas and virtues.
Not to mention dozens of paper books, which will be humanity’s last incorruptible witnesses—the first enemies of super-intelligent artificial intelligences. A couple of years ago, I wrote a novel about attacks on neighborhood libraries and the persecution of these paper beings as subversive merchandise. If I haven’t made faster progress on this novel, it’s not just because fiction always takes (or takes me) much longer than nonfiction—and because it often ends in self-destruction. It’s because today’s culture has already succeeded in making its authors begin to question the purpose of their own work. After every book, I tell myself, “This was the last one,” and although I’ve systematically broken that promise, I feel closer and closer to that moment with each one. Only our frustrations with the world push us to do it again, but everything (as in Jorge Luis Borges’s wonderful poem, Limits) has a “no more.”
The writings I referred to at the beginning are texts that originated 100 percent from my keyboard. However, they are now assessed as having 10 or 15 percent AI generation. The answer (or hypothesis) is paradoxical: naturally, AI systems will show a certain percentage that matches their writing styles and content for the simple reason that they’ve been plagiarizing us for years.
It’s no secret. Reports like the one in The Atlantic include thousands of authors, among whom I find myself. In other databases, a dozen books appear. The self-reference is because, coincidentally, I am the author I know best—though not completely. In the case shown in the screenshot accompanying this text, the books on the list are only those that have come to light due to lawsuits filed by other authors. The AI systems themselves acknowledge that this indicates the list is much longer. On the other hand, I don’t even want to know what they’ve done with the 900 articles I’ve published openly, without an ISBN or copyright.
Of course, in this case, the purpose has always been to disseminate ideas and to examine social and cultural issues. That’s not my problem. It’s not that some of those 900 articles—which are also part of an ideological and cultural ecosystem—are being repurposed. That’s how social thought has functioned for millennia. Nor am I concerned that the AIs process them without even mentioning all their original sources—an endemic weakness of AIs. In fact, I have written in articles and in the introduction to some of my latest books that, for some time now, we have been writing intentionally for AIs, since readers are increasingly turning to them to interpret books and texts of all kinds—with the inherent risks of the AIs’ ideological biases and conceptual errors or misinterpretations of facts.
Another paradox arises with AI-generated opinions. Their “politically correct” responses (i.e., those that do not upset those in power) fall apart if “consumers” are not satisfied with the first or second response and engage in a debate with any of them. As we’ve noted in other articles, not only is it easy to win an argument against them, but that doesn’t matter when millions of people believe their first response as if they were impartial and omniscient gods. The paradox arises when we see that, for example, in 2024, AI systems like ChatGPT insisted on historical interpretations that we openly contradicted in books and classrooms—such as the reasons for Texas’s independence— and today they repeat those same conclusions from the outset, so students are no longer surprised or scandalized by our theories, nor do newspaper readers call us conspiracy theorists in the comments section, as they used to.
Well, that’s one of the problems with plagiarizing us. Writers, artists, and scientists in general—and throughout history—tend to hold positions opposed to the narratives of power, which is why, in a world dominated by capitalism and fascism, they are often discredited as “leftists,” “left-wingers,” or “indoctrinators.” However, true indoctrination (implanting a doctrine in someone) can only be carried out on people lacking the intellectual resources to defend themselves—that is, (1) idiots or (2) children. The first group is debatable, because no one accepts that label for themselves, but the second is objective: children lack the defenses against indoctrination. Taking them to a church, making them believe in Santa Claus, in millionaire geniuses, or in the wickedness of the poor—like their parents—is pure and simple indoctrination. We professors work with young adults; children go to churches or are educated through screens owned by the ideological partners of the churches of Mammon. And while universities have the power to shape public opinion, they will never surpass—in either quantity or quality—the influence of churches, media propaganda, or the political and financial sermons of the super-millionaires. Because a belief will always be stronger, more popular, and easier to consume than an idea supported by verifiable facts and substantiated arguments—all of which require moral and intellectual effort; and because power, even when concentrated in the hands of a handful of men, always needs the support of millions who hate their enslaved brothers and sisters and defend their masters, who present themselves as the self-sacrificing victims of yesteryear.
It is to be expected that AIs will succeed in correcting this weakness of power—which arises when its own instruments of domination come dangerously close to truths independent of economic, financial, political, media, religious, and military powers.
My problem isn’t even the use of AI in our work. It’s a broader issue: AI isn’t owned by the people who created it; it was hijacked (privatized) just a few years ago. That not only means that concentrated capital has once again hijacked humanity’s new inventions, but that it is using them to dominate, exploit, and enslave that very humanity. All without even asking permission. All thanks to the magical power of having amassed excess capital before multiplying theft and its total dictatorship.
After centuries of enrichment through slavery, in the 19th century the empires accelerated the colonization and destruction of Africa by reinvesting their surplus capital. Today, the world is Africa, yet we perceive ourselves as free white people.
Jorge Majfud, 10 de agosto de 2026
Deja una respuesta