Recent redundancies and office closures have sharpened the focus on the impact of Generative AI (GenAI) on journalism. Business Insider recently laid off 21% of its staff, citing a slowdown in traffic to its site. This was followed by the apparent closure of TechCrunch's European operations.

These examples are indicative of a deeper challenge within journalism and scientific research—the upheaval caused by Large Language Models (LLMs). OpenAI, Perplexity, Anthropic, and others are impacting traffic to news portals and altering the way news is curated and produced. While these challenges should not be underestimated, there is also potential for GenAI to be a force for good. Harnessed correctly, it can create new revenue streams and enable a greater focus on investigative journalism and research.

A GenAI-induced upheaval

At the beginning of 2025, the Reuters Institute released its report on the health of journalism. It made for stark reading. Of the editors and CEOs surveyed, only 41% had confidence in journalism's future, and 75% were concerned by the drop in traffic to publishers' sites.

Falling traffic is a growing problem. Traffic from social media has been in steady decline since 2022, although some publishers have seen an uptick in traffic from Facebook in recent months. It is the loss of traffic from search engines, however, that is really concerning publishers. Apple recently reported its first-ever recorded drop in Google search queries on its Safari browser.

The emergence of GenAI is largely responsible for this drop. According to a recent survey, 61% of Gen Z and 53% of millennials are choosing to search the web using chatbots like ChatGPT. The issue for publishers is that AI chatbots provide readers with a summarised version of the content, which makes going to the source unnecessary in many users' eyes. The result is lower traffic and, therefore, lower advertising revenue for publishers.

Google embraces GenAI

As the world's leading search engine, Google was initially reluctant to embrace GenAI for search. However, last Summer, it changed tack with the release of Google AI Overview, which provides users with an AI-generated summary along with the traditional list of websites. Google has since doubled down on AI, introducing AI Mode for Google search and adding ads to this feature a month later.

Google is keen to emphasize that the increased integration of AI into its search does not necessarily mean less traffic to publishers. Liz Reid, Head of Search at Google, pointed out that "links included in AI Overviews get more clicks than if the page had appeared as a traditional web listing." However, this assertion stands in contrast to the message from most publishers, including Business Insider, who stated their recent staff cutbacks were the result of lower traffic levels.

Concerns for journalists

With Google now fully in the AI game, concerns for the health of journalism and research abound. The revenue loss caused by reduced traffic to publishers is certainly the largest concern. While some of the biggest players can look to subscriptions for the bulk of their revenue, for many smaller and more niche publications, running ads on their site is their most consistent and effective source of revenue. And advertisers will only pay you if you have traffic to your site.

However, search traffic is not the only potential problem GenAI is causing. AI summaries risk oversimplifying complex issues. As Joshua Rothman, a staff writer at the New Yorker, recently commented, "A reporter might labour for months to unearth new information, only for A.I. to hoover it up and fold it into some bland summary." These concerns are noteworthy, but so are some considerations from a more optimistic perspective.

Different business models and a return to in-depth journalism

GenAI's disruption is not one-dimensional, and for every challenge it is creating, there are also potential new benefits.

Firstly, the GenAI shift could support a return to higher-quality research and journalism. The advertising-based business model employed by many publishers has generated revenue. However, it has also encouraged clickbait and led outlets to focus on quick-turnaround summaries rather than in-depth investigations. The result is parallel coverage, where hundreds of publishers all report on the same events in the same manner.

A shift away from this model could see journalism return to a more investigative and in-depth approach. Encouragingly, there is evidence that this type of output may be in high demand in the era of GenAI. PR professionals have noted that LLMs appear to prioritize high-quality content from credible publishers when answering search requests. This includes small-scale media, such as local newspapers and trade publications. While it is not yet obvious how publishers and journalists might monetize this opportunity, it is a positive sign that quality journalism will still have an important role to play in the AI-dominated information landscape.

Naturally, publishers need to survive financially for this to be possible. And here, GenAI may be able to compensate for the revenue losses it is causing. Specifically, publishers are looking to earn from their output by licensing it to GenAI companies. For example, the New York Times has just signed a deal with Amazon to license its content for LLM training.

New approaches to research and reporting

The smart use of GenAI by publishers, researchers and journalists could also help new formats to thrive. For example, GenAI can help publishers provide an increasingly personalized news experience, a trend noted as of significant importance in the Reuters Institute report.

GenAI tools can also be utilised to handle repetitive tasks, freeing up human talent for more investigative and in-depth work. Furthermore, the granular research required for this kind of output can be supported by AI and other tech tools. Zach Seward, Editorial Director of AI at the New York Times__, describes the potential for__ "investigations that involve tens of thousands of pages of unorganized documents, or hundreds of hours of video, or every federal court filing."

This powerful application of GenAI can be paired with advanced web scraping that automates large-scale data collection from the web. Web scraping already supports high-quality investigative journalism and research through initiatives like Project 4β, which provides free access to these tools. GenAI can help sift through data scraped from the web and identify noteworthy facts or patterns.

Between the two scenarios

How the advancement of GenAI will actually play out for journalism remains to be seen in the coming years. The worst-case scenario is publishers scaling back in the face of lost traffic and journalists resorting to ever more precarious freelance work to survive. While AI-generated content may fill some of the gaps, continued issues around trust, provenance, and accuracy may create deep mistrust across the information ecosystem.

The best-case scenario is one where GenAI drives new forms of revenue to publishers, especially through licensing. In this case, journalists and researchers will enhance their status through the role they play in generative search results, while new AI tools and greater access to data will help them produce quality work. The result is a return to in-depth investigative journalism that provides nuance and insight to complement AI-generated summaries and overviews. For this scenario to materialize, organizations must support journalists and researchers in their pivot toward deep, data-led and AI-assisted journalism.