Artificial intelligence is turning into a tool that is becoming increasingly powerful, easy to use, and popular. It also brings growing concerns and dangers, nearly in parallel: from cybersecurity to data protection, misinformation or dehumanization.
The most recognized and analyzed critique of an improper application of AI in our lives has been that of Pope Leo XIV. In Magnifica humanitas, he manages to address the major pillars of the difficulties we face when trying to control this technology. It is something comparable to the industrial revolution, only even more massive. Of course, in this era the media are, or we are, involved, and I do not think we are meeting what one would expect of the fourth estate.
A few days ago, while reading an interesting analysis that aligned with the criticisms Leo XIV made of AI, I realized that the text itself was written, for the most part, using some kind of AI agent. It was signed by a person with a long career at one of the country’s major outlets. Yet, there wasn’t even a note that they were relying, precisely, on the very tool they vehemently criticized, which made their arguments illegitimate. Unfortunately, this is just one of several occasions on which these practices have been detected.
“AI, when used properly, is a great support for analysts and newsrooms because it streamlines tedious and mechanical processes”
Even the pontiff himself has been accused of writing his encyclical with AI—even if this is false—. If the media are required to deliver immediacy and cover all possible informational spaces, they will seek out the tools to try. AI, when used properly, is a great support for analysts and newsrooms because it streamlines tedious and mechanical processes, leaving more time for creativity, strategy and asking good questions. However, the way it is being used today only achieves the speed our time demands at the cost of a terrible drop in the quality of what we deliver to readers. It is not desirable for an agent to write our text from start to finish, even if there is a revision involved. If we do, what prevents readers from going directly to that information from the AI itself? What is the value of journalism if it cannot offer something different from what anyone could obtain using Claude or ChatGPT? I’m not a journalist, but I think it’s necessary that we ask ourselves these questions.
Many of the problems arising from this excessive and non-transparent use have been highlighted by the Journalism Laboratory of the Luca de Tena Foundation. Drawing on data from a recent study, they note that “texts classified as AI-generated or AI-assisted showed an average semantic similarity 33% higher than texts not generated by AI”. In addition, there are other biases, such as a tendency toward optimistic analyses — which is 107% higher compared to texts written by humans — or the public’s own distrust. 75% of respondents in this study say that AI-generated content tends to contain more “factually incorrect information”.
At the same time that AI is being used indiscriminately by the media, it poses an existential threat to our own business model. Today it is not enough to be the fastest or to bring the best analysis, because, even in that case, our content risks being eclipsed.
Previously, editorial rooms paid thousands of euros to try to understand how Google’s algorithm worked and which SEO techniques would allow a given piece of news to come first when certain keywords were searched. However, today we are witnessing a sharp and rapid shift in the way readers approach the media. Instead of searching the internet, it is increasingly common for one to rely on their trusted AI to find the information.
“The same tool that is driving the media’s hyper-productivity is the one destroying its profits and its reputation”
During the process, all outlets lose visits and visibility. In other words, money. If the reader does not go to your page, they don’t see your ads or cross your paywalls, you’re losing money. It is ironic that the same tool driving media hyper-productivity is the one destroying its profits and reputation. It is also favorable for those who advocate understanding AI differently, recognizing the destructive capacity of creativity and imposing limits and rules while seeking consensus. Because what today lets you compete with others, tomorrow can sink you.
In this regard, a somewhat more demanding reader would also be desirable. Let us reward original content —from humans to humans— and develop a greater critical sense. Everything moves faster today and escaping this dynamic is almost impossible, but we could lay the first stone on the path if, instead of asking AI for a summary of any text over eight hundred words, we devote those two extra minutes to reading what lies before us.
“If AI has not come, among other things, to fix these problems nor to give us more quality in our content, what is it here for and who benefits?”
We face an enormous challenge. Perhaps the greatest of our time. And yes, there’s no doubt that the media will have to deal with AI one way or another, but we must pause to think. Yesterday, El Mundo, Marca, and Expansión published all their articles without bylines as part of the strike their workers began last May. They report fifteen years of salary stagnation. If AI has not come, among other things, to fix these problems or to improve the quality of our content, what is it here for and who benefits?
