AI revolution: how neural networks threaten the survival of traditional media

This significantly reduces advertising revenue for media outlets that are already facing tough times, Techxplore reports.

"The next few years are going to be incredibly challenging for all publishers without exception," warns Matt Carolian, vice president of research at Boston Globe Media.

He believes publishers must proactively adapt or risk disappearing under the onslaught of AI-generated summaries and summaries.

A recent Pew Research Center study confirms media experts' fears: users are much less likely to click through to primary source sites if a search engine gives them a short AI-generated response to a query in advance. The number of clicks on links in the presence of such answers is roughly halved.

For media outlets already struggling to survive, this is a serious blow: traffic has a direct impact not only on advertising revenue, but also on the number of new subscriptions. John Wiebe, a professor at Northeastern University in the United States, believes that these changes will soon radically transform the Internet environment.

Technology giants Google and Meta have long been cutting media revenues from advertising, forcing publishers to shift to a paid subscription model. But subscriptions, too, depend on an influx of new users. And subscriptions alone, as Wiebe emphasises, cannot fully sustain large media companies.

Some publishers, like the Boston Globe, are trying to adapt by using neural networks to their advantage. But the number of new subscribers attracted through AI is still negligible. Other popular tools, such as Perplexity, have yielded minimal subscription gains at all.

In response to this trend, a new strategy - GEO ("generative neural network optimisation") - has emerged. It involves creating clear and high-quality content that is easy to analyse by AI systems. However, a fundamental question arises here: should AI services be allowed to freely use media materials at all?

Many publishers are strongly opposed to this approach, blocking AI crawlers from accessing their content. Daniel Coffey of industry association News/Media Alliance believes that if AI companies want to use media content, they should pay a fair market price.

Some media outlets have already partnered with big tech companies. For example, the New York Times has signed an agreement with Amazon, Google is partnering with the Associated Press, and French news agency AFP is partnering with AI developer Mistral. However, the conflict is far from over: high-profile lawsuits have already been filed against OpenAI and Microsoft, the most high-profile of which is from the New York Times.

Publishers are faced with a difficult choice: by closing content for neural networks, they protect their rights, but deprive themselves of potential readers. As a result, more and more companies are returning access to AI crawlers in an attempt to find a balance.

However, even that doesn't guarantee success. According to OtterlyAI, media accounts for just 29% of all sources cited by ChatGPT. By comparison, corporate websites make up 36%. And while Google traditionally gives priority to recognised media, ChatGPT and similar systems often don't, making it harder for publishers.

The problem goes beyond economics and touches on audience trust. According to a Reuters Institute report, already about 15 per cent of young people under 25 get their news through generative neural networks. Given the difficulty of verifying the sources of information from such systems, the risk of disorientation and fake news increases significantly.

"Someone has to do the original journalism. Without original material, AI services simply have nothing to analyse," Karolian stresses.

Google has already started to enter into partnerships with the media, perhaps recognising the need for cooperation. However, according to Professor Wiebe, it is unclear whether publishers will be able to adapt before the consequences become irreversible.