Laura Turini is a lawyer specializing in intellectual property, she curates the LISP newsletter on Substack and often writes for Appunti. How have Italian publishers reacted in these first years to the arrival of artificial intelligence? Has it really changed the business model or just the context?
At the moment, it seems to me that mainly the context has changed, while the business model appears substantially unchanged, with publishers clinging to models that now seem obsolete.
Certainly, artificial intelligence has been welcomed within newsrooms to facilitate ordinary work, even though the Journalists’ Association has imposed precise obligations, updating the code of ethics: on one hand the duty to verify news, on the other the obligation to clearly indicate when a piece has been written with the help of artificial intelligence systems.
So, on one side it is well understood that one cannot give up this new tool, which facilitates work; on the other, its dangers are highlighted.
The greatest danger, from a commercial point of view, is represented by AI Overviews: that system whereby artificial intelligence intercepts news, summarizes it, and offers it already digested to users, who no longer need to consult the original page, resulting in reduced traffic and advertising revenue.
The publishing industry is reacting to this phenomenon in two ways. Some try to get captured by AI Overviews, in order to gain at least some visibility. Others react with opt-out, that is, introducing systems to prevent AI Overviews from acquiring their news, in an attempt to then be able to enter into licensing agreements and derive income from them.
These are two opposite ways of facing the problem: one adaptive, the other more defensive. But neither, at the moment, seems particularly effective to me.
The feeling is that publishing is somewhat groping in the dark, also because no one expected such a rapid acceleration of this new technology.
Based on what is observed in other sectors, what is realistic to expect regarding the impact of AI and especially artificial agents in the media world?
Autonomous agents are increasingly widespread in many technological sectors, and this suggests that sooner or later they will also arrive within newsrooms.
Compared to the traditional generative artificial intelligence we know, an agent is a much more complex system: it uses artificial intelligence but also connects to other tools that allow it to acquire information and news and then act in the external world.
In practice, an agent can search for information on the web, read, write, send emails, query databases, execute parts of code and even write it. It can therefore completely replace a human being in some specific functions for which it is programmed.
Some newsrooms, especially in the United States, are already using agents. For example, Bloomberg and Reuters have been using automated systems for years to produce standardized financial and sports news, which are among the easiest to generate. However, it seems they now want to use them also for more structured news.
These systems are also employed to personalize feeds in real time and provide news to a fragmented audience, rather than a generalist one.
It is recent news that Mark Zuckerberg himself is designing his own personalized agent to help him make strategic decisions. So it is easy to think that, sooner or later, an agent capable of replacing a journalist, at least at a medium level, will be designed and created.
On the legal front, the issue of the origin of training data remains open. But even here, unfortunately, it can be hypothesized that platforms will prevail, acquiring entire newsrooms more as data assets than as editorial assets.
According to current estimates, a hope for journalism remains in quality journalism: investigations, reports, contextual analyses, which will maintain their differential value. But they will probably be aimed at a smaller group of users willing to pay.
For all others, a great attention challenge will open up, because it will be up to readers to distinguish true news from false ones improperly generated by artificial intelligence.
(Excerpt from Appunti by Stefano Feltri)
