Artificial intelligence is driving a fundamental shift in the news environment

Artificial intelligence is leading to a fundamental change of the news environment
Artificial intelligence is shaking up journalism and will cause a "fundamental change in the news landscape" in the short term, media expert David Caswell told AFP. As a former employee at Yahoo! and BBC News Labs, the UK broadcaster's innovation arm, Caswell spoke to industry leaders gathered in the Italian city of Perugia to discuss the biggest issues facing their sector.

How do you see the journalism of the future?

"We don't know. But what we're trying to do is understand all the possibilities, or as many of the possibilities as we can. But I think there are some things that are becoming clearer: one is the fact that more media is likely to be created and generated by machines. In this way, machines will gather more information in journalism, do more with production, with audio, video and text, and create the kind of consumption that will appeal to consumers. This is a very fundamental change in the information environment in general and in the news environment in particular. This is structurally different from where we are now. We don't know how long it will take - it could be two, four, seven years. I think it will be faster because there is very little drag. People will not need news devices, new hardware, they as producers will not need a lot of money, technical expertise. All these things that were barriers in the previous generation of artificial intelligence (AI) are no longer barriers thanks to generative AI."

What are the latest events in the newsrooms?

“One development vector is new tools that allow the AI workflow, for example at Politikens in Denmark, to focus on making existing products and activities more efficient. But it is also the basis for the transition of their products, their workforce and their activities in this new environment. Or in the world of AI. There's a tool that Google has created -- codenamed "Genesis" -- that's being tested with publishers. Some publishers build their own. There will be cross-platform versions of these tools. These are tools where you bring your news collection to the left side: files, transcripts, audios, videos,... for example. This helps you do things like analysis, summarizing, converting to screenshots, audio recordings. They are orchestrated by the instrument. What the journalist does is coordinate the tool, check the content through and edit. The work becomes using the tool as an editorial manager, which is the fruit of AI. Technically this works. But it's quite a different thing to run it in a newsroom, in a large newsroom, and use it day after day, month after month. That's a big question: Will it be embraced with enthusiasm, will it be used in a way that isn't very productive in the long run, or will it dramatically increase newsroom productivity?''

What is the price?

"In the last decade, it was very expensive. It was very difficult: you needed the data, you had to build a data warehouse, you had to do a business deal with Amazon cloud or Google Cloud, you had to hire data specialists, you had to have a team of data engineers. It's a big investment. Only the BBC, the New York Times, or media outlets of that caliber could afford it.

This will not be the case with generative AI. You can manage the news workflow through interfaces for which you pay $20 per month. You don't have to be a programmer. All you need is motivation, enthusiasm and curiosity. There are a lot of people in news organizations who haven't dealt with AI in the past because they didn't have the technical background, but now they're just going to be able to use it. It's a much more open form of AI: just as smaller edits can do a lot, younger individuals in more established edits can do a lot. I think that's a good thing, but it's also a destructive thing. Often the internal policy in the newsrooms is violated by this.

What stage of AI are we at?

“AI has been around since the 1950s. But AI for practical purposes came with ChatGPT. It will be a long time - years - before we really figure out how to use them for valuable things. There are so many things you can do with them. The risk to journalism is that other organizations, start-ups, technology companies will do things in the news faster than the news world itself. Many startups have no editorial component at all. They steal content from news organizations, some cover niches: monitor press releases, social media channels, PDFs of reports.

What are the risks?

"Journalism hasn't done well in the last 10 or 15 years, there's really no credible vision of the future of how it's going to play out just in the social media world." What AI does (is) it gives news organizations a chance to change that situation, to participate in a new system. It's good to be optimistic, to engage, to explore, to do projects, to experiment, maybe to change your way of thinking, that's positive. As Jelani Cobb, dean of the Columbia School of Journalism, says, "+AI is an overwhelming force around which journalism will have to organize itself." There is no way artificial intelligence can adapt to journalism. /BGNES, AFP