How Information Organizations must redo their activities as the AI general threatens their existence

Business news

March 18, 2025

Hi and welcome to AI. In this edition … The media fought with the II; Trump prescribes us efforts to reorient the fight against “ideological prejudice”; Distributed learning acquires an increasing craving; More and more powerful II can tilt the scale to totalitarianism.

AI potentially destroys business models of many organizations. However, in several sectors, the threat would seem existential as the news. It’s like mine, so I will, so I hope you forgive a somewhat indulgent newsletter. But the news must be important for all of us, since the functional free press plays a significant role in democracy – the formation of the public and helping to prosecute power. And there are some similarities between the way the news executives are not critical – they refuse the problems and opportunities that AI presents, which can be learned by business managers in other sectors.
Last week, I held a day at the Aspen Institute conference called “AI & News: Charting Cournation”, held at Reuters’s Kokvter in London. The conference was attended by the top heads of a number of British and European information organizations. It passed according to Chatham House rules, so I can’t tell you who said exactly but I can convey what was said.

Tools for journalists and editors

The news executives talked about using the II, first of all, in domestic products to make their teams more effective. AI helps write, optimized search engines and translate content, allowing organizations reach a new audience They did not traditionally serve in places, although most emphasized that people were kept in the cycle to monitor precision.

One editor described using AI for automatic production of short articles from the press -release, releasing journalists for more original reporting while maintaining human editors for quality control. Journalists also use the II to generalize the documents and analysis of large sets of data – as state documents and satellite drawings – interacting with journalistic journalism, which will be difficult without these instruments. These are cases of good use, but they lead to modest impact – basically, making existing workflows more effective.

Bottom up or top to bottom?

Among the Heads and Technical Sciences, there was an active discussion on whether information organizations should accept the approach to the bottom-up-conducting generative tools of AI in the hands of each journalist and editor, allowing these people to manage their own data analysis or “Viba” Widgets running on AI to help them at work or whether it is worth the top -down effort and prioritize projects prioritizing.
The bottom-up approach has merit-one democratizing AI access, enables Frontline staff, who often know pain and can often notice good use before the high-level performers can, and releases AI developers to spend only on projects that are big, more sophisticated and potentially more strategically important.
The disadvantage of the from the bottom up is that it can be chaotic, which impedes the organization to ensure the preservation of ethical and legal policy. It can create technical debt if tools are built on the go, which cannot be easily maintained and updated. One editor is worried to create a bunk edition, and some editors cover new technologies and others lag behind. The bottom up also does not guarantee that the solutions generate the best profitability of investment-elimination of AI models can be quickly more expensive. Many called for a balanced approach, though there was no consensus on how to reach it. From the conversations I had with the performers in other sectors, this dilemma is familiar in various fields.

Cautious about the danger of trust

News outfits are also cautious about creating AI audience tools. Many started using the II to resume articles that can help busy and increasingly impatient readers. Some built ai chatbots that can answer questions about a particular, narrow subgroup of their coverage – such as stories about the Olympics Either climate change– But they usually call them “experiments” to help pull off the readers that the answers may not always be accurate. Few people went on in terms of AI-generated content. They are worried that the hallucinations producing the gene will be inappropriate to the accuracy of their journalism. Their brands and their business eventually depend on this trust.

Those who hesitate will be lost?

This caution, though clear, itself is a tremendous risk. If the information organizations themselves do not use AI to generalize the news and make it more interactive, technology companies. People are increasingly resorting to AI and chats search engines, including surprise, Openai Chat, as well as Google Gemini and “AI” reviews now provides in response to numerous searches and more. Several news executives said at the conference that “disisterMediation” – a loss of direct connection with their audience – was their greatest fear.
They have reasons to worry. A lot of information organizations (including Fortuna) At least partially depend on Google search to attract the audience. A recent Tollbit Study– The selling software that helps to protect web -rates from web -coroners – it is up to the Google AI review speed was 91% lower than the traditional Google search. . Clickwhich also offers help protect the publishers of news from scraping, found this Openai scraped the news site 250 times for each direction page.

So far, information organizations have responded to this potentially existential threat through a combination of legal deviation – The New York Times sued Openai for copyright violations, and Dow Jones and NEW YORK POST sued the surprise– And partnerships. These partnerships participated in long -term, seven -digit licensed news transactions. (Wealth My partnership is both surprised and with Prorata. They also saw affiliate relationships as a way to establish relationships with technology companies and press their experience to help them create AI products or teach their staff. No one saw the relationship as particularly stable. All of them knew about the risk of becoming excessively dependent on the income from AI licensing, burning earlier when the media -industry allowed Facebook to become the main engine of traffic and advertising revenue. Later the money disappeared almost for the night when Meta Following the US presidential election in 2016, CEO Mark Zuckerberg decided to depleting the news in the folk channels.

Ferrari who works on AI

The executives recognized the need to build direct audience relationships that could not be hung with AI companies, but the few had clear strategies for this. One expert at the conference directly said that “the news industry does not take the II seriously”, focusing on “additional adaptation rather than a structural transformation.” He compared modern approaches to the three -stage process that had “Ferrari” at both ends, but “horse and basket in the middle”.

He and another advisor on the media industry urged information organizations to get away from the structure of their approach to news around “articles”. Instead, they urged news artists to think about ways of source (public data, interview transcripts, documents obtained from sources, raw video materials, audio recordings and archival news) can be transformed into different output-forming videos, short-term videos, bullet resumes, or so, traditional Technology. They also urged information organizations to stop thinking about news production as a linear process, and start thinking more about it as a round loop, perhaps the one in which there was no man in the middle.

One person at the conference stated that information organizations should become less islands and look more closely at the understanding and lessons of other industries and how they adapted to the II. Others said that this may require startups – perhaps the information organizations themselves – for pioneers of new businessmen for AI age.

The rates could not be higher. While AI creates existential problems for traditional journalism, it also offers unprecedented opportunities to expand the achievement and potentially connect with the audience that “shut off the news” – when the leaders are bold enough to reconsider what news can be in the AI era.

This is still the news of AI.

Jeremy Kan
jeremy.kahn@fortune.com
@jeremyakahn
Correction: Tuesday on Tuesday on Tuesday on AI Improperly identified country where TrustPilot is with the headquarters. This is Denmark. In addition, the news in this edition incorrectly identified the name of the Chinese startup for the Viral AI Mastor Manus. The startup name is the butterfly effect.

Originally this story was presented on Fortune.com