In the era of hyper-fake, hyper-personalization and online disinformation, artificial intelligence (AI) poses crucial challenges to the veracity and integrity of the information disseminated. Media efforts to address it are “primitive,” according to the conclusions of various experts drawn at the Collision conference, held last June.
The media is not approaching AI in the right way, says Julien Toyer, editor-in-chief of the Thomson Reuters news agency. Speaking at the Collision technology conference in Toronto, the editor-in-chief noted that a lot of the questions are about the future of journalism and whether it will disappear, which “takes us away from really focusing on the audience, what we need to do. This tool can actually provide us with very powerful solutions to some of the audience problems that we’ve seen in recent years.”
More than ever, major media outlets are struggling to capture readers’ attention, says Matthew Kaminski, editor-in-chief of Politico. “Time is the most precious thing we all have. How do we not waste our readers’ time? How do we reclaim their time? These new questions are one of the biggest changes that newspapers have seen in the last 25 years. I think it will only accelerate with AI.”
“Primitive” AI efforts
Despite this acceleration, the editor-in-chief of Politico sees three interesting applications of AI for media: fostering the development of a better relationship with the reader, improving their experience of journalistic content and enabling newsrooms to be more efficient at lower cost.
Mr. Kaminski reveals that the American media is training an AI to “write headlines in the style of Politico “, summarises articles and translates texts. He would like to see more newsletters designed according to the reader’s interests. However, he considers that, for the moment, the media’s efforts in terms of AI are very “primitive”.
It’s not new that media outlets are struggling to adapt to new technologies, notes CBC News reporter Ellen Mauro. While nearly a third of Americans under 30 regularly get their news on TikTok, there’s still limited journalism created specifically for the platform. Instead, mainstream outlets tend to create news stories in more traditional ways first and then adapt them for TikTok, she notes.
“But I think we need to try to get to a point where this transformation of content will be more natural and will be more part of the media’s procedures. Right now, people want authenticity from us, because they feel they get it from citizen journalists who simply publish videos,” explains the journalist.
Mme Mauro believes that this reluctance to move to new platforms parallels the media’s tendency to be cautious about AI. “As usual, when it comes to technology, we [les médias] “We’re a little late and caught off guard. The internet and search engines have been very disruptive, which leads the media to think that AI would be highly destructive,” adds Matthew Kaminski, who acknowledges that the media’s lack of financial resources also hinders AI adoption.
The advent of the Internet has contributed to the mass closure of newspapers in the past two decades. Since 2005, nearly 2,900 U.S. media outlets have closed their doors, according to a report from Northwestern University’s Medill School of Journalism, Media, Communications and Marketing. Nearly two-thirds of journalists are estimated to have lost their jobs during that same period, representing about 43,000 positions.
The risks of hyper-personalization and hyper-faking
Muhammad Lila, founder of Goodable, calls the rapid development of AI a “very dangerous time for media.” He continues: “This technology brings the possibility of hyper-personalization, and we all like to consume content that fits our preconceived notions.”
A 2021 research published in the Journal of Quantitative Description: Digital Media reveals that 10% of Americans rely solely on partisan news sources.
Lila worries that AI will increasingly lock readers into echo chambers and that conversations between people with divergent opinions will be filled with misinformation. “Eyeballs used to be the ultimate level of verification. Now we’re getting to the point where that’s no longer a barrier to truth. We’re actually entering a period that’s sometimes called a post-truth world,” the speaker says.
In 2023, 43 per cent of Canadians said it was “more difficult than three years ago to distinguish between true and false information,” according to a Statistics Canada survey. Lila says these risks are growing as the sophistication of deep fakery blurs the lines between fact and fiction.
“Now what happens when you are faced with images that are impossible to verify, but appear to be real? […] To me, that’s an incredibly scary thing, because truth has always been a fundamental layer that unites us. And if you take that away, what do we have left?”
AI to “enrich” the reader experience
Despite the fears surrounding AI, speakers still cast a positive eye on the technological revolution in the news. “I think it breaks down barriers, makes people easier to reach, and allows you to take the audience with you everywhere,” says journalist Ellen Mauro.
For Mr. Lila, technology offers the public a voice. “Our job as storytellers is sometimes to step aside and let people tell their own stories. And that’s the best thing that technology will allow us to do: let people tell their own stories.”
AI can also become an enabler of original content produced by journalists, according to Matthew Kaminski. “The experience on a website is much richer than in a newspaper, so I can only imagine that what comes after that will be much richer. I think it should improve our journalism and the technological revolution has put an emphasis on originality.”