<\/p>
Bartosz W\u0119glarczyk recently asked a direct question: \u201cWhy would anyone type in our address directly?\u201d Writing in September, Onet’s editor-in-chief<\/mark><\/a> described the decline of a model in which media outlets focused primarily on reach allocated by algorithms. He argued that individual journalists, original stories and reporting should give readers a reason to return.<\/p> W\u0119glarczyk is pointing to a basic principle of journalism: readers need to know why they value a particular newsroom. For years, online publishers could make money without that attachment. A user simply had to arrive at an article through a search engine, read it and see an ad. They didn’t even have to remember the site’s name. Now Google can answer a question using someone else’s publication and keep the user on its own platform. The newsroom still bears the cost of producing the article, but its opportunity to earn revenue from it is becoming less certain.<\/p> An answer without a visit<\/strong><\/p> Users don’t need convincing of the benefits. A few sentences in search results save time, let us skip the ads and give us what we came online for. When the answer is enough, opening the source becomes an extra step. A link beneath the summary does little to change that.<\/p> Pew Research Center examined the behavior of 900 American adults<\/mark><\/a> using Google in March 2025. When the search engine displayed an AI summary, users clicked a traditional result in 8% of visits. Without a summary, the figure was 15%. Links within the AI answer itself were opened in just 1% of visits. In this group, identifying the source rarely meant that anyone actually visited it.<\/p> Publishers are already feeling the loss of readers arriving through search. The Reuters Institute cited Chartbeat data for more than 2,500 websites<\/mark><\/a>, showing a 33% decline in traffic from Google’s organic search results between November 2024 and November 2025. It’s unclear how much of that decline was due to AI Overviews. These were the global figures W\u0119glarczyk referred to. They shouldn’t be read as Onet’s own results.<\/p> The threat is easiest to see in articles answering simple, recurring questions. If their entire value to the reader can be condensed into three sentences, a search engine has little work to do to replace a website visit. Worse, the findings of an expensive investigation can be summarized in much the same way. The length of an answer says nothing about the work involved in establishing the facts.<\/p> Polish websites count their page views<\/strong><\/p> Polish data also gives publishers cause for concern, though it doesn’t justify attributing the entire loss to AI. Wirtualne Media compared the performance of 130 websites using<\/mark> Mediapanel data<\/a>. In July 2026, they recorded a combined total of about 5 billion page views, down from 6.26 billion a year earlier. That was a 20.1% decline, while their combined reach, with duplicate users counted only once, fell by just 0.5%. The sites were still reaching almost as many people, but those people generated substantially fewer page views.<\/p> However, those figures reflect different stories. Polska Press attributed much of the decline at naszemiasto.pl to a change in the way galleries were displayed, after which a single visit no longer generated as many page views. Meanwhile, a PBI analysis published in January 2026<\/mark><\/a> didn’t indicate a sharp collapse in search traffic to Poland’s largest websites. The datasets cover different periods and measure different things. There’s no basis for treating every lost page view as a reader taken away by AI.<\/p> Nor are all publishers equally dependent on Google. In November 2025, Jacek \u015awiderski said<\/mark><\/a> the search engine accounted for just over 10% of page views across WP’s sites, while more than 80% came from direct visits. The head of Wirtualna Polska acknowledged that Google traffic was falling but believed the company could offset the losses through other efforts. He also highlighted the more difficult position of smaller publishers, which depend more heavily on platforms.<\/p> \u015awiderski’s and W\u0119glarczyk’s comments should be read together. The former described the advantage of having an audience that remembers a website without a search engine’s help. The latter asked how to earn that place in readers’ minds. A large, well-known brand has an advantage here. Simply telling a smaller newsroom, previously discovered mainly through Google, to build an equally loyal audience isn’t enough. It still has to fund its operations until it gets there.<\/p> Publishers have their own bill to settle<\/strong><\/p> The industry would find it easier to defend its role if it hadn’t spent years making it harder for readers to get what they came for. A headline promises an answer, but the article takes a long time to deliver it. Readers have to find the key information among the introduction, ads and more paragraphs that add little. Against that backdrop, an AI answer’s brevity may be its greatest strength.<\/p> Rafa\u0142 Madajczak raised this uncomfortable issue for publishers in a response to W\u0119glarczyk published on September 26 in Wirtualne Media<\/mark><\/a>. The former Gazeta.pl editor-in-chief pointed to the pursuit of reach, clickbait headlines and the tendency to bury answers. He called on websites to be useful to their audiences. A newsroom’s belief that it produces valuable content won’t, on its own, make readers seek it out.<\/p> Publishers can’t expect readers to click out of a sense of duty to the media industry. They need to give them a good reason: information they can’t find elsewhere, a writer they trust or an explanation that truly improves their understanding of events. And they shouldn’t force readers to wade through ads to get there.<\/p> That doesn’t absolve platforms of responsibility for the terms under which they use other people’s work. Even an excellent, easily accessible article can be summarized in search results. Improving publication quality is a newsroom’s responsibility, but it doesn’t settle who should earn money from the subsequent use of that work or on what terms.<\/p> How much should Google pay?<\/strong><\/p> Polish publishers have long been trying to establish those terms. On August 4, the Digital Publishers’ Employers’ Association (ZPWC) announced<\/mark><\/a> that mediation with Google had ended without an agreement. It said it would ask the president of Poland’s Office of Electronic Communications (UKE) to set the terms for the use of press publications and the level of compensation. The dispute is based on legislation implementing Article 15 of the DSM Directive and concerns the use of press content beyond AI answers alone.<\/p> The sums involved are substantial. One model presented in a FehrAdvice report commissioned by ZPWC<\/mark><\/a> estimated that Polish media outlets should receive 514 million PLN a year. That figure was commissioned by one party to the dispute. It isn’t a court-ordered payment or a total of AI-related losses. Publishers are using it to put a value on the benefits Google derives from including their publications in its services.<\/p> Going to court doesn’t guarantee payment either. As Reuters reported on October 1, a US court dismissed antitrust lawsuits brought by Chegg and Penske Media against Google<\/mark><\/a> over AI Overviews. Those rulings don’t amount to blanket permission for AI to use content however it likes. But lost traffic alone wasn’t enough for the antitrust claims presented to succeed.<\/p> Google is also testing direct payments to publishers for their publications’ contribution to AI answers. The \u201cAI contribution pilot,\u201d described by Digiday in September<\/mark><\/a>, covers selected partners. Payment depends on whether Google considers a piece of content a significant contribution to a generated answer. Some of the people Digiday spoke to welcomed the opportunity to work together, while others complained about low payments and an unclear calculation method. The company confirmed that this was an early pilot.<\/p> Under this arrangement, publishers still have little say. One company decides whether to use a publication, how to assess its importance and how much to pay for it. Revenue may come in, but it’s hard to plan a newsroom’s budget around a valuation whose rules are unknown. Small websites also need a way to get paid, even if they lack the bargaining power of a large media group.<\/p>