NEW YORK– Generative AI search assistants such as ChatGPT, Grok 3 and Gemini are starting to take over some of the space once held by traditional online search. This shift is pulling audiences away from news outlets and news media, which are already working hard to stay afloat as advertising earnings drop.
Matt Karolian, who leads research and development at Boston Globe Media, expects tougher times ahead. He believes the next few years will be tough for the news media, with AI summaries reducing the number of visitors to news media sites. “News Media must prepare, or they risk being left behind,” he said.
Recent data support these concerns. According to a Pew Research Center report, when Google adds AI-generated summaries to search results, people are less likely to click through to the actual articles from news media.
Users are twice as likely to skip the source links compared to traditional search results. For digital news outlets that rely on visitor numbers for ad sales and subscriptions, this drop-off is a real threat.
John Wihbey, a professor at Northeastern University, says this trend is only speeding up. Before long, web traffic and news delivery could look very different from today.
Big Tech News Media Help
Big tech companies like Google and Meta have already hurt news media companies by scooping up most of the ad money, making publishers shift focus to paid subscriptions. Wihbey points out that even paying readers depend on ongoing traffic, and subscriptions alone can’t keep large newsrooms going.
Boston Globe Media has found that some new subscribers come through ChatGPT , giving them a new way to reach readers, Karolian noted. Yet, these gains are small compared to other sources, including less popular search engines.
Other AI tools, such as Perplexity, bring even fewer subscribers, making them less valuable for news media.
To adapt, publishers have begun using Generative Engine Optimization (GEO). This method builds on traditional SEO by making content easy for AI models to read, with clear labels, good structure, plain language, and strong visibility on social sites and forums like Reddit.
A key issue remains:Should news outlets let AI crawlers scan their articles? Thomas Peham, CEO at OtterlyAI , says many publishers now block these bots to protect their work after past cases of aggressive data collection by large AI companies.
OpenAI and Microsoft Sued
Danielle Coffey, who leads the News/Media Alliance, argues that news companies deserve fair payment for their content. Some progress is being made through licensing deals, such as agreements between the New York Times and Amazon, Google and AP, or Mistral and Agence France-Presse.
But major legal fights are ongoing, including the New York Times suing OpenAI and Microsoft . Publishers must choose: block AI and protect their work, or allow access to attract more readers. Peham has seen more publishers reconsider and reopen their content to AI crawlers, hoping to reach bigger audiences. Yet, even with open access, results vary.
OtterlyAI analysis shows news outlets account for only 29% of sources cited by ChatGPT, while business websites take the lead at 36%. Google search results often favour trusted news outlets, but ChatGPT does not always follow the same pattern.
This issue isn’t just about business. Reuters Institute reports that about 15% of those under 25 are now turning to AI to read the news. Since some AI tools do not always show their sources clearly, this shift could leave many unsure where their information comes from or how reliable it is, much as happened with social media.
Karolian stresses that original journalism is still at the heart of news, saying, “Without real reporting, AI platforms would have little to summarise.”
Google is already working with news media outlets to improve how its AI features use news media content, hinting at possible cooperation in the future. Wihbey believes that platforms will eventually recognize their need for the press, but whether this awareness comes quickly enough for struggling publishers remains to be seen.
