Google upgrades Search with Gemini 2.0 and introduces AI Mode
Google just integrated Gemini 2.0 into AI Overviews in the U.S. to tackle coding and advanced math, while dropping the login requirement and expanding access to teens. The search giant also launched 'AI Mode' in Labs, a deep-reasoning search experience that runs concurrent multi-topic queries and supports conversational follow-ups. Early access is currently rolling out to Google One AI Premium subscribers.

Searching the web for nuanced topics usually means drowning in twenty open browser tabs, cross-referencing specs, and piecing together answers by hand. Google wants its search engine to do all that heavy lifting for you.
Google announced two major moves for Search: an upgrade that brings Gemini 2.0 to AI Overviews, and a brand-new experimental experience in Labs called AI Mode.
Overviews get Gemini 2.0 and lose the login wall
AI Overviews have already crossed one billion users worldwide. Now, Google is swapping in a faster, more capable brain.
In the U.S., AI Overviews are now powered by Gemini 2.0. The upgrade focuses on tougher queries, starting with programming, advanced mathematics, and multimodal searches. Responses should be quicker, higher quality, and triggered more often for technical topics.
Google is also tearing down access barriers. Teens can now view AI Overviews, and users do not even need to sign into a Google account to see them. It is clear that Google treats AI summaries as the default front door of search, not a niche perk.
Under the hood of AI Mode: Query fan-out
For power users who find regular summaries too shallow, Google rolled out an early Labs test dubbed AI Mode, powered by a custom version of Gemini 2.0.
Instead of treating a prompt like a single ping, AI Mode acts like a multi-step research agent. It uses an approach Google calls query fan-out, where the model breaks down a complex question into multiple subtopics and fires off concurrent searches across the web, the Knowledge Graph, real-world data, and shopping records covering billions of products.
If you ask to compare sleep tracking between a smart ring, a smartwatch, and an under-mattress mat, AI Mode builds a plan, runs simultaneous searches, refines its findings on the fly, and returns a unified breakdown. Users can then ask continuous follow-up questions to drill deeper.
Google noted that if the system lacks high confidence in factual accuracy, it will fallback to standard web links rather than guessing. Even so, Google cautioned that early tests might occasionally reflect unintended personas or opinions. Access is rolling out first to Google One AI Premium subscribers.
Why it matters
If you spend hours researching electronics, business ideas, or complex concepts, AI Mode could save serious time. It basically automates the dirty work of querying multiple angles and synthesizing findings. That said, gatekeeping early access behind a paid Google One AI Premium subscription means everyday searchers will have to wait.
We think Google is playing directly to its unfair advantage here. Competitors like Perplexity and OpenAI have forced Google to rethink its interface, but neither possesses Google's Knowledge Graph or its massive real-time commercial shopping index. Hooking multi-step agent reasoning into those proprietary data layers makes this much more than a standard chatbot wrapper.
However, the strategy brings serious risks. If AI Mode answers complex questions thoroughly in a single pane, it chips away at the remaining referral clicks sent to publishers and independent blogs. There is also the question of compounding errors: when an AI coordinates several sub-searches behind the scenes, catching a hallucinated premise buried in a synthesized answer becomes that much harder for users.
Sources (1)
- blog.google(opens in a new tab)Originalblog.google
Topic trackerGoogle AI Mode
- Google’s AI Mode gets eyes
- Google upgrades Search with Gemini 2.0 and introduces AI ModeThis story