Loading article…
Google adds AI Mode to Search, using Gemini 2.0 for deeper reasoning and multimodal answers; over a billion users already see AI Overviews.
Google rolled out AI Mode, an experimental Search feature that pairs a custom Gemini 2.0 model with its Knowledge Graph to answer complex, multi‑part queries and surface follow‑up links, expanding the AI Overviews experience that already reaches more than a billion users [2].
| At a glance | |
|---|---|
| Feature | AI Mode in Google Search |
| Model | Gemini 2.0 (custom version) |
| Availability | Limited, opt‑in Labs rollout to Google One AI Premium subscribers |
| Reach | AI Overviews used by > 1 billion people worldwide |
The new AI Mode builds on the AI Overviews product, which Google says is already used by over a billion people for quick answers. Gemini 2.0 powers the mode, delivering “faster and higher‑quality responses” for hard queries such as coding, advanced math, and multimodal requests [2]. Unlike standard search, AI Mode runs a “query fan‑out” that issues multiple related searches across subtopics, then stitches the results together, giving users a single, conversational answer with links for deeper digging.
AI Overviews previously offered single‑step answers; AI Mode adds multi‑step reasoning, follow‑up capability, and access to fresh, real‑time sources like the Knowledge Graph and shopping data. Google notes that testers praised the speed, quality, and freshness of responses, but the company also warns that the system may occasionally generate answers that sound opinionated or “take on a persona” [2]. The rollout is currently limited to Google One AI Premium subscribers, with broader Labs testing planned.
Google’s AI Mode mirrors emerging generative‑AI search features from rivals such as Microsoft’s Bing Chat, which also blends large language models with web indexing. However, Google’s integration of its own Knowledge Graph and the “query fan‑out” technique aims to provide broader coverage than a single‑model approach. The company’s broader AI experiments—including Gemini‑powered tools like Google Flow, NotebookLM, and the Nano Banana 2 image model—show a coordinated push to embed generative AI across its product suite [1][4].
AI Mode signals Google’s intent to make generative AI a default layer of Search, turning what was once a supplemental overview into a full conversational experience. Whether the approach can consistently outperform rivals while maintaining factual accuracy remains to be seen.
Coverage is mostly measured — 120 of 132 reports stay neutral.
Every Monday — the token unlocks, Fed dates & catalysts set to move crypto and markets this week. So you’re never blindsided.
Free · 3-min read · one-click unsubscribe
AI-assisted synthesis by the TrendWatcher Editorial Desk · sourced from 4 outlets · Jul 3, 2026 · How we report
Google AI merged its internal Google Brain unit with DeepMind in 2023 to create a unified entity called Google DeepMind.
Gemini is Google's AI assistant, formerly known as Bard, available through subscription plans (AI Plus, Pro, Ultra) in over 140 countries with usage limits based on computation.
Google AI has research locations in Zurich, Paris, Israel, and Beijing.
Alphabet removed a guideline that prohibited using its AI technology for applications likely to cause harm.
Google AI contributed TensorFlow, a machine learning library, and Magenta, a deep learning project for creative applications.