Google searches are forever changing. Enter the “AI agents” and that comes with a wave of changes and even concern, for some

For more than two decades, Google Search worked on the same simple logic: you typed a few words into a white box, and the search engine returned a list of links to sites that might contain the answer. This model has defined the way people have navigated the Internet since the early 2000s. Now, Google is explicitly saying that this era is coming to an end.

At the annual Google I/O conference, the company unveiled the biggest transformation of its search engine in 25 years.

The central idea is that Google no longer wants to be just an intermediary that sends users to other sites, but a space where answers, explanations and even actions happen directly inside Google Search, with the help of artificial intelligence.

Basically, less Google Search, more AI chatbot.

“Search Google” becomes “ask Google's AI”. But the answer is not simple

Instead of just getting a list of results, if you choose “AI Mode,” Google will generate summaries, interactive explanations, graphs, widgets, and even “visual experiences built in real time” based on the question asked.

If you ask about black holes, for example, Google could directly show you an interactive simulation or visual representation that you can explore and modify with additional questions, according to Liz Reid, Google's head of search.

The company also wants to change the way people formulate searches. Google encourages longer, more natural and conversational questions. Instead of typing in a few keywords like “Bucharest weekend weather”, Google says you should ask questions like “Is it worth going on a trip to the mountains this weekend if it rains on Saturday?”.

The role of “AI agents”

One of the most important changes is the introduction of so-called “information agents”, i.e. AI agents that can search and monitor information instead of the user. Basically, Google proposes that you delegate part of the documentation activity to artificial intelligence. You can set such an agent to constantly monitor certain topics, changes or trends online and send you summarized updates when something relevant comes up.

Google gives the example of financial markets: a user could ask an AI agent to track fluctuations in a certain economic sector according to very specific criteria. The agent would decide for itself which sources and data to track, review the changes, and then come back with a summary and some additional links to dig deeper.

Journalist Vlad Dumitrescu sends the Good Tech newsletter every Wednesday morning. If you want to receive practical tools to make your life easier with the help of technology, you can subscribe here:

People can make their own “mini-apps”

In the new model envisioned by Google, the links are no longer the central point of the experience, but rather become an appendix for those who want to go deeper into the subject.

Google is also taking an important step in another direction: turning the search engine into a platform where people can build small custom apps. The company says people will be able to create their own “mini apps” in Search tailored to their needs. For example, someone could create a meal planning app using information from their personal calendar or a fitness program tailored to their goals.

All of these features are powered by Gemini, Google's family of AI models, and an internal platform called Antigravity, dedicated to developing interactive AI agents and interfaces. Google says the new experiences will be available for free to all users starting this summer.

How Google's changes translate concretely

The stakes are huge. If until now Google was like a kind of gateway to the Internet, the new direction shows that the tech giant wants people to spend more and more time directly inside the Google ecosystem, without having to actually access the sites that produce the information.

For decades, the Internet was built on a relatively simple trade-off: websites produced content, and Google sent traffic to them. The press, blogs, forums, online stores accepted to depend on Google precisely because the search engine brought them readers, users and implicitly income.

Of course, people can always opt out of the “AI Mode” option. But why would they? Already, many people have started looking for information directly in chatbot conversations instead of searching on Google. People run away from any form of friction if they come across it and will always choose comfort.

If the answers are generated directly in Google, if AI summaries synthesize the information before the user even enters the source site, and if AI agents do the documentation instead of humans, then an increasing part of the traffic risks simply disappearing. Of course, Google will still display links in AI Mode answers, but only at the end of a paragraph so you can see where the information came from.

Anxiety among publishers

In fact, the phenomenon has already started with the introduction of AI Overviews, the summaries automatically generated by Google on top of the classic results. In the last year, many sites have noticed significant drops in traffic from Search, because users receive their answer directly in Google and have no reason to click further. Now, the company is basically saying this is the future.

We must not forget that Google still needs the content produced by sites to feed these AI responses. The company's models are trained on human-created text, articles, forums, videos and discussions published online. It's just that the economic model that supported the production of this content is starting to fracture. If publishers are losing traffic and money, who will invest in the long-term production of original information that AI summarizes and reassembles?

There is already real anxiety in the media industry and in the ecosystem of independent creators that the Internet could enter a vicious circle: AI consumes and recomposes content created by humans, but at the same time reduces the visibility and income of those who produce it. In the long run, this could lead to a poorer online environment dominated by content optimized for algorithms and AI models, not real readers.

This shift also says a lot about how the Internet itself is transforming. For years, the online environment has been built around the idea of ​​exploration: compare sources, join forums, read obscure blogs, or discover new sites.

Now, the big AI companies are pushing the Internet toward a model where information comes already processed, synthesized, and packaged into a single interface.