The AI ​​bubble may burst. People will not use AI for information. » Nieman Journalism Lab

When it comes to AI, there are growing whispers that we are in bubble territory, perhaps headed for a burst next year. The main reasons commonly cited are the wide gap between spending on AI computing infrastructure and AI revenue, and the lack of a clear business model for some companies. When a bubble bursts, which is by no means a given, it's not a good situation for investors, for the state, or for some of the companies that have been driving the development of recent years. Expect market consolidation, a lot of hassle, and “hey, I told you so!” from the most annoying people online.

But here comes a bit of a problem, especially when it comes to news. Even if we do see such a market correction, the technology won't magically disappear, nor will many of the features and products built using it, nor will the fact that people increasingly embrace it to understand the world around them.

First, not all AI companies offering direct-to-consumer AI capabilities are the same. OpenAI could struggle to generate revenue to pay for its investments (and partner debt) if the going gets tough. Attracting the hundreds of millions of paying members needed will not be easy. The advertising market is already highly saturated, and corporate consolidation is similarly competitive. The headwind is strong. But consider Google in contrast. They have huge financial resources and are leading in search. A growing cloud business offsets the risk on the advertising side. The company trains its models on its own chips, has its own world-class research department, and is leading the way in developing the models. As Melissa Heikkilä succinctly writes in the FT, “Google can combine its dominant positions in search, cloud infrastructure, and smartphones to bring new AI capabilities to billions of existing users.” What she doesn't say: This is probably here to stay, even if big AI ships start pitching. Even in a post-bubble world, some of the companies that transformed the news and information landscape will continue to shape our digital landscape.

Second, despite the sometimes barely concealed vitriol of some experts when it comes to AI as an information gateway, public AI adoption and attitudes paint a different picture. People generally seem to prefer having access to a greater variety of information. They also like convenience, speed, and specificity. AI systems often provide all three, not only when it comes to chatbots, but also when AI is integrated into search and other digital products. So perhaps it's no surprise that 43% of people expect AI to improve their experience interacting with search engines, compared to 17% who said the opposite (for social media, it was 33%/15%, for news, 26%/30%…).

Third, more and more people believe that AI, either alone or on its own platform, can help them navigate the complexities of life, although not always in a perfect way. Our research shows that the use of new AI systems and tools to impact information discovery is rapidly increasing, with information retrieval now the primary use case. The UK's Institute for AI Security found something similar regarding political intelligence during the 2024 UK elections, while the Pew Research Center shows that the use of AI “to learn new things” is increasing in the US. Similarly, Similarweb's 2025 report shows increased use for these and related purposes. This doesn't mean that people blindly trust (in fact, don't believe) these systems, or that they don't want them to be managed properly, but it's very different from the lamentations you'll hear from some of the chattering classes.

So my predictions are: In 2026, the use of AI as an information gateway will further increase. It would be unwise for news organizations to hope that the bursting of the AI ​​bubble will stop that trend.

Instead, they will have to contend with the fact that even if those systems and features do not directly compete with news, competition for attention, advertising, and the money that people and organizations are willing to spend will only intensify.

For the news industry, this means the difficult task of figuring out where it fits in this strange new world continues. This is neither impossible nor unsolvable. The best thing about news media is that it not only quenches our thirst for knowledge, but also has many things that people still value today, such as depth, empathy, nuance, originality, surprise, and humanity, which are provided by AI. But the challenge of leveraging these strengths will not get any easier in 2026, with or without sonic boom-level “props” as the AI ​​bubble bursts.