AI Search vs Google: Will Classic Web Search Shrink by 2030?

A paradox has emerged online in 2026. ChatGPT, Gemini and other AI assistants are already mass-market tools for finding information, yet Google Search is not showing an obvious collapse. Google instead says total query volume has reached an all-time high as AI Mode expands.

The contradiction disappears once we separate the need to search from the classic search interface. What may decline is not the number of questions people ask, but the share of journeys in which they get a list of links, open several websites and assemble the answer themselves.

ORBK.NET’s current estimate is that classic web search has about an 80% probability of losing a substantial share of informational queries to AI assistants and generative search interfaces by the end of 2030, with a working range of 75–85%.

This does not mean Google loses 80% of the market. The more likely transformation is that the search engine itself becomes an AI assistant.

People are already searching differently

According to Pew Research, 42% of U.S. adults said in February 2026 that they use AI chatbots to search for information. That is no longer a niche behavior.

At the end of August 2026, OpenAI said ChatGPT was available to more than 1 billion weekly active users. This is a company-reported figure rather than an independent measure of search-market share, but it demonstrates the scale of asking an assistant directly.

Google is not trying to freeze the old search-results page. In May 2026 the company said AI Mode had surpassed one billion monthly users and that total Search queries had reached an all-time high.

The key signal is simple: AI can take activity away from classic search without taking it away from Google as a company.

What counts as a substantial loss?

Classic web search means an interface where the user submits a query, receives primarily a list of web results and then moves between pages to construct the answer.

AI search includes ChatGPT, Gemini, Claude, Perplexity, Copilot and other assistants, plus generative modes inside traditional search engines when the primary output is a synthesized answer or conversation.

The forecast defines a substantial shift as at least 25% of total informational search activity. AI does not merely need to be popular; it must replace roughly a quarter of the work that would previously have been done through classic search.

AI’s strongest advantage: it removes a step

Classic search mostly finds documents. An AI assistant tries to produce the answer. That shortens the chain from question to outcome: search, open pages, read, compare and synthesize.

Ask why a washing machine is not draining and a traditional engine may return a forum, the manufacturer’s manual, videos and repair sites. An AI system can compare those materials, highlight likely causes and ask a follow-up question.

The difference becomes larger in conversation. After asking which laptop to buy for video editing, the user can refine budget, software, weight and battery life without launching a new series of searches. The real competitive advantage is fewer actions between the question and a useful answer.

The impact on websites is already clearer than the impact on search engines

A Pew behavioral study using data from 900 U.S. adults and 68,879 Google searches in March 2025 found that users clicked a traditional result in 8% of visits when an AI summary appeared, versus 15% when no AI summary appeared. Links inside the AI summary were clicked in only about 1% of visits.

This is an observed association, not proof that AI caused the full difference. But the direction matters: a user may get enough information from the search page to skip the source website.

For Google, that can still be a successful search session. For a publisher, it is a lost visit. One of the first major casualties of AI search may therefore be the old traffic-distribution model.

Classic search still has a strong defense

Users do not always want a synthesized answer. Official documents, specific websites, stores, tickets, breaking news and disputed claims often benefit from direct access to sources.

Trust is another limit. AI can misunderstand a request, blend sources or produce a confident but wrong answer. That matters in law, health, finance and current events.

A Gartner survey of 377 U.S. consumers found that only about one-third considered generative chatbots as effective as search engines for learning new information. Thirty-one percent said AI summaries made them spend more time searching, compared with 16% who said they spent less.

That is an important counterargument. AI may not reduce total search at all. It may add a new loop: ask the assistant, verify in Google, open a primary source and return to the assistant.

Google can win even in a post-classic-Google world

Google already controls one of the strongest distribution channels for AI: Search itself. It does not need to move billions of users to a new website; it can redesign the interface inside the product they already use.

In 2030, people may still say “I’ll Google it.” But after the query, the system may conduct several searches, read pages, compare sources, ask follow-ups and take the next action.

In that world, the YES forecast can resolve even if Google retains a very large market share. “Will Google lose search?” and “will classic web search lose informational queries?” are different questions.

History provides a mechanism, not a clean base rate

Generative AI is too new, and the category of informational search is changing alongside the technology. Inventing a numerical historical success rate would create false precision.

A more useful pattern is that new information interfaces often do not destroy the previous channel; they force it to absorb the new function. Here, AI assistants will not only replace search from the outside — search engines themselves will become AI assistants.

Three scenarios through 2030

What would change the forecast?

The estimate should rise if independent measurements in 2027–2028 show sustained growth in the share of informational queries handled by AI assistants; if AI Mode becomes Google’s default interface for complex queries; if research, shopping and planning workflows move systematically into assistants; and if citation and primary-source handling improve materially.

The estimate should fall if chatbot usage stabilizes mainly around writing, coding and work tasks; if users almost always return to classic search after an AI answer; if generative answers remain too expensive or unreliable; or if legal and market restrictions sharply reduce AI access to the open web.

The strongest signal is not ChatGPT’s user count itself. It is which tasks users stop doing through classic search.

What this means for websites

SEO is unlikely to disappear, but its goal may change. A site will compete not only for a click but also for its data, brand and expertise to be selected inside an AI-generated answer.

That creates a paradox: AI needs high-quality websites in order to find and synthesize information, but answer-first interfaces can reduce the traffic that finances those sites. The future of search is therefore also a question of who funds the open web if users increasingly do not need to open the underlying pages.

Classic search probably will not die — it will stop being classic

People will still search for websites, official documents, stores, maps, videos, news and primary sources in 2030. But for “explain,” “compare,” “recommend” and “research” tasks, a list of links is no longer obviously the most convenient interface.

The most likely outcome is that classic web search does lose a substantial share of informational queries by 2030, but much of that loss is captured by the AI version of search itself.

Forecast card

Forecast disclaimer: This forecast does not claim that the event will occur. It is a current probability estimate based on information available at the time and may change as new information appears.