This interview is part of a series of conversations with creatives and professionals who deal, now every day, with the advantages and limits of artificial intelligence
Vittorio Roidi has seen a lot of journalism: he worked at Radio Rai, at Messaggero, he was president of the Fnsi, the Federation of the Italian Press, secretary of the Order of Journalists and sign Writing and journalism ethics at the School of Perugia. You’ve seen a lot but perhaps never like what you’re seeing now. So much so that she felt the need to put it in a book, which she calls herself The Artificial Journalist and is published by Edizioni All Around.
The title is a clear clue on the topic: How does artificial intelligence enter the profession of those who provide information? How can it improve it? How can it make it worse, if not destroy it? And can he really destroy it? We asked him.
Artificial intelligence The New York Times sues OpenAI and Microsoft for violating Emanuele Capone’s copyright December 27, 2023
Roidi isn’t sure that AI can destroy the journalist’s workbut without a doubt “we must start to be alarmed, we can’t just write about it, talk about it, talk about it – they told us when we reached him on the phone – We must also act and move”.
There is a precise moment in which he began to do so, meaning he began to become alarmed: “I read a piece of Rampini (this, ed.) in which he openly admitted that he had lost the writing competition with ChatGPT after asking OpenAI’s AI to write an analysis on one of the topics he knows best, the East”. In his article, Rampini actually says that ChatGPT he accomplished that task “in five minutes”providing a result that is “dignified not only in terms of form, spelling and syntax”, but “also in terms of content: a balanced and updated synthesis of information on the topic”.
In short, to put it in Roidi’s words: “If even someone like Rampini, with his experience, recognizes it, then there is something good in these AIs”. And yet, there is also something worrying: “Honestly, you can’t hear those who say all technological revolutions caused initial upheavals but then things settled down.” Why? “Because if someone like Geoffrey Hinton leaves Googleif the same people who develop these tools they say we should be carefulthen maybe we really should be careful.”
But we as journalists, why should we be alarmed?
Because there is a concrete risk of further discrediting (the word used is another, ed.) of a category of which people already do not have a very high opinion. And we must move above all for two reasons: before more layoffs arrive because we will be replaced by machines and before we pick up other bad habits in addition to those we already have, such as working less, working badly and working more comfortably.
And what path should we follow to defend ourselves?
First of all, the rules must be established, which must be there and are fundamental: the EU has put its own through the AI Act, the Fnsi has asked for other common sense ones, such as that any use, even minimal, is made explicit to the reader , of an AI in composing an article. And the Journalists’ Association rightly pointed out that this topic should be included in the next national employment contract, for example by establishing that every piece that comes out, online or on paper, must be checked by a human hand.
We need rules, like those obtained by Hollywood actors, but AI does make some positive contribution to our work, right?
Sure. There are great opportunities even in the daily routine and wonderful things can be done if we do them well.
Which things?
For example, these tools prove useful for helping you with long, tedious or difficult tasks, such as processing and putting in order the results of a complex survey or even having a 30-line crocodile made (a piece about someone’s death, ed.) that you didn’t have ready if a famous person dies at midnight and you need it for the next morning’s paper. There ChatGPT can definitely speed up everything, and do the job in a few minutes instead of an hour. Although the journalist must always remember to verify and above all to check the most recent information, because these tools are not always updated on current events.
Does this also apply to images?
This is very true for images, because now even with photos the risk is enormous: once upon a time, if you recovered a photo of a fact or the protagonist of an event, if one of your sources gave you an image, you could be at peace. It was a test. Not now: it’s no longer like that. But this is another field where technology can help us a lot in verifying information.
The cover of The Artificial Journalist (image from ufficioreporter.eu) Change! The photojournalist Fabio Bucciarelli: “As long as a man takes photos in Gaza, AI will not scare me” by Pier Luigi Pisa 15 January 2024
And in the future? What will our work be like in 10 years?
We will have increasingly perfect and precise software, machines and AI, but whoever wants to do this job will still have to know how to write. Well made, well thought out and cared for pieces will always be useful. Publishers need to understand that they need to hire good journalists, and they need to pay them well, if they want to sell their products and stories and be successful. And they have to make fewer cuts and less advertising.
In what sense, less advertising?
Unfortunately, there are already examples of low or very low level information: it happens when publishers use journalists to push a topic dear to them, advertise something, perhaps another product of theirs, editorial or otherwise. As I said, this causes the entire category to lose credibility, and AI and its dangers fit into this context. Risking amplifying these problems and further widening the distance between the writer and the reader.
Speaking of the reader, what is good information for readers?
It is often said that “journalism will be saved by quality”, but are artificial intelligences really quality? AIs don’t make mistakes, they are precise on lengths, they are fast, they respect deadlines and don’t make errors, but are they capable of providing quality information? Am I able to make original contributions? I do not believe. Instead, the reader really needs quality, explanations, mediation, even more so in these complex times, as demonstrated by the Covid two-year period, in which it was difficult to understand what was happening and the role of this type of journalism has once again become central .
However, AI could also be a source of income for publishers, perhaps by establishing agreements for the use of their content: could this be a solution?
I don’t know if it could be a solution, but I know that publishers should think less about profit, defend those who do this job and be more attentive to the fact that in every country democracy also comes from good information. Bad information produces bad voters, as demonstrated by the case of Donald Trump in the United States, with the flood of fake news that he spread during the election campaign that saw him become president.
In short, to summarize: AI they are undoubtedly a useful tool also in journalism but “we cannot wait and be overwhelmed”, as objectively happened with the arrival of the Internet in Italy between the mid and late nineties, because “the world is changing, and we we can’t stand by and watch”. Nor can we remain passive.
@capoema