po polsku
illustration: DALL-EGoogle has been dealing the cards in the attention game for years, but in 2024, it changed the rules - to its advantage. The launch of AI Overview, automated summaries generated by artificial intelligence, gave the zero-click search phenomenon a whole new dimension. It’s no longer just quick answers to quick questions. Now it`s full-featured articles shown without visiting any website.
We’ve seen this before. ChatGPT, Perplexity and other AI models are right behind the giant, offering fast, concise answers without digging through dozens of links. It’s a new quality in information searching - faster, simpler, and less distracting. To keep users from leaving, Google is now playing the same card. Convenient? Yes. But is it fair?
AI Overview. A new face of no-click searching
What is zero-click search? It’s when a search engine gives you the answer on a silver platter before you even click anything. You type in a question, and the engine shows a definition, a conversion, a map, or - more and more - a full summary. Google no longer settles for showing a snippet. Now it builds out the full answer. AI Overview is a module that generates a summary of a user’s question based on multiple sources. No click needed. The user gets what they came for - and the content publisher gets nothing.
Rand Fishkin from SparkToro warned last summer that out of every 1000 searches on Google, only 360 clicks in the EU and 374 in the US go outside Google’s ecosystem.
| Google Searches 2024 | USA (%) | EU (%) |
|---|---|---|
| Total Clicks | 41.5% | 40.3% |
| No Action (Session Ends) | 37.1% | 37.4% |
| New Search | 21.4% | 22.3% |
| Zero-click searches | 58.5% | 59.7% |
This isn’t just about technology. It’s a paradigm shift. Information is no longer tied to its source. Even when Google shows links, they’re subtle - and users often don’t feel the need to visit. Because they feel they "already know".
Classic zero-click was already a problem. But AI Overview raises the stakes. It removes the user from the whole clicking process. Before they even look through the results, they’re given a ready-made dose of knowledge. No visits, no sessions, no revenue for publishers. It’s like running a restaurant where Google sends people, shows them a photo of the menu, lets them smell the food - but doesn’t let them in. In an industry where every click is a potential customer or reader, that’s like cutting off the air supply.
New SEO. Optimizing for AI, not humans
We used to fight for the top spot in search results. Now the goal has shifted - you need to show up in the AI summary field. The problem is, the rules aren’t clear. Google doesn’t say what determines which sources are chosen. That means traditional SEO is no longer enough.
We have to write content not only for users, but in a way the AI algorithm can "understand". Think structure, clarity, definitions, and direct answers. The more "textbook-like" your content appears, the more likely it gets into AI Overview. Is this still journalism - or just content farming?
AI Overview doesn’t match keywords directly. It analyzes the user query as a whole and identifies intent. That’s why your texts need to be not just accurate, but contextual. You need to anticipate what the reader is really looking for - even if they can’t phrase it clearly.
This is a big change for content creators. It’s not enough to write "how to cook pasta". You need to know if they want cooking time, calorie count, or gluten-free options. AI will figure it out. If you don’t, your content disappears. Zachariasz Kijak, SEO expert at digital agency Verseo, explains it well in the guide How to optimize for LLMs.
Vanishing brand and AI mistakes
When Google creates AI Overviews, it pulls from specific sites. But users rarely know which. Sometimes there’s a link, sometimes a name - often, nothing. The content is "generated", as if it created itself. But someone wrote it. Someone is responsible. Someone wants credit. This isn’t just an ethical issue - it’s a business problem. Because:
- if users don’t see the brand, they build no loyalty;
- they won’t return or remember;
- and that kills brand communication through content.
Worse still, AI can do a lot - but it’s not flawless. AI Overviews sometimes include mistakes, misrepresentations, or entirely made-up facts. Yet to the average user, it’s still a "Google answer" - something trustworthy.
7 facts about news on social media 👇
That’s the trap. When AI is the source, there’s no clear accountability. If an article contains an error - you know who wrote it. When AI gets a summary wrong, no one feels responsible. But the consequences can be serious - from bad health advice to political misinformation.
Who’s really getting paid?
At the end of the chain lies the question that’s being asked more and more: who should profit from content? Google - for displaying it? Or the author - for creating it? AI Overview is a perfect example of a model where the middleman profits, and the creator gets nothing.
Publishers are starting to push back. There are calls for licensing fees, revenue sharing, new partnership models. If Google uses others’ work to increase ad profits - fairness demands that it shares.
If AI Overview becomes the standard - and it looks like it will - the internet will undergo a major shift. On one side: user convenience. On the other: the death of organic traffic as we know it. And the question the industry keeps asking: do we really want an internet where Google eats the whole cake - and leaves everyone else the crumbs?
COMMERCIAL BREAK
New articles in section Media industry
Information laundering. Russian propaganda returns as news from China
KFi
Russian intelligence agencies use complex tactics to mask the primary source of their harmful narratives. Patrycja Krzyśpiak from NASK details this dangerous process of information laundering in a recent report.
Health disinformation on the internet. Diagnosis from an algorithm
Katarzyna Czechowicz
Health disinformation can cause a cancer patient to see a doctor when it is too late for effective treatment. It can also undermine trust in doctors and public institutions. As a result, people may refuse to cooperate even during a real external threat, experts warn.
Feminatives are not a given. How Polish women talk about their professions
SWPS
Polish women still rarely use female job titles. This comes from the latest report by the Center for Research on Social Relations at SWPS University. Women's professional activity proved to be a significant factor in using feminatives.
See articles on a similar topic:
How artificial intelligence misrepresents the news. PBC analysis
Sylwia Markowska
In news summaries generated by the most popular models in Polish, as many as 46% of responses contained at least one significant error, 27% had serious issues with sources (missing, misleading, or incorrect), and 19% contained hallucinations and outdated information.
Social Media and Relationships. Interesting Research from Palestine
KFi
What does love look like in the digital age? Does technology bring people closer or push them apart? In an era where Facebook and Instagram replace dinner table conversations, social media has become a new space for marital relationships. Researchers from An-Najah National University examined how technology can build bonds but also sow uncertainty.
YouTube vs. Television. The 50+ Generation Shifts to Computers
Krzysztof Fiedorek
For years, so-called "silvers" were primarily associated with traditional media like television. However, research by IQS for SilverTV and Lifetube shows that this view is outdated. The report’s findings clearly demonstrate that YouTube is becoming the new “television.”
Yellow Badge. Jan Bluz's documentary on political prisoners in Belarus
BARD
Imagine writing three posts on a social networking site. For a few clicks on a keyboard, you get three years in a penal colony. Sounds like a grim joke? For political prisoners in Belarus, this is the reality that Jan Bluz shows in the documentary "Yellow Badge", produced with the support of the Pulitzer Center.
