Google has launched support for five new languages in its AI Mode search feature, marking a significant step toward making advanced AI tools accessible to billions of non-English speakers.
Announced on September 8, 2025, the update adds Hindi, Indonesian, Japanese, Korean, and Brazilian Portuguese, allowing users in over 180 countries to interact with the platform in their native tongues.
From English-Only Experiment to Global Powerhouse
AI Mode debuted in March 2025 as an opt-in experiment within Google Search Labs, initially limited to English and available only to U.S. subscribers of Google One AI Premium.
It quickly evolved into a full rollout in the U.S. by May 2025, following announcements at Google I/O.
The feature expanded to India and the UK shortly after, and by August 2025, it reached 180 countries in English.
This multilingual expansion is the first major language update, addressing the English-only barrier that had limited its reach despite Google's long history of supporting over 100 languages in standard search.
Early adoption has been strong, with reports of a 10% increase in complex query volume in English markets.
The rollout timeline highlights Google's rapid iteration:
-
March 2025: U.S. launch via Labs, powered by Gemini 2.0.
-
May 2025: Full U.S. access; international testing begins.
-
June-July 2025: India and UK rollouts; voice and image features added.
-
August 2025: Global English expansion to 180+ countries; agentic tools like reservations introduced.
-
September 8, 2025: Five new languages added via Gemini 2.5.
Why Now? Competition and User Demand Drive Change
The timing reflects intensifying rivalry from AI search platforms like Perplexity and OpenAI's ChatGPT Search.
Google aims to retain dominance in non-English markets, where over 5 billion people live.
Hema Budaraju, Google's Vice President of Product Management for Search, emphasized that true global search requires more than translation—it demands cultural nuance and local relevance.
Gemini 2.5 Powers Multilingual AI Capabilities
At the core is a customized Gemini 2.5 model, Google's most advanced "thinking" AI, released in March 2025.
This multimodal system processes text, images, voice, and video, using a "query fan-out" technique to break complex questions into subtopics and search across Google's Knowledge Graph, real-time web data, and billions of product listings.
For the new languages, Gemini 2.5 is fine-tuned for local contexts, handling scripts like Devanagari for Hindi and incorporating regional dialects, idioms, and cultural references.
Key capabilities include:
-
Advanced Reasoning: Excels in math (86.7% on AIME 2025), science (84% on GPQA), and coding (63.8% on SWE-Bench), outperforming rivals in benchmarks like LMSYS Arena.
-
Multimodal Inputs: Users can upload images or use voice for queries like health advice in Korean or event recommendations in Brazilian Portuguese.
-
Local Nuance: Avoids literal translations; for example, it suggests Brazil-specific festivals or Korean financial tools tailored to user preferences.
-
Context Window: Handles up to 1 million tokens, with plans to expand to 2 million for deeper analysis.
Agentic Features & Personalization
AI Mode now includes agentic tools for tasks like restaurant bookings via partners such as OpenTable and Resy, though initially limited to U.S. Google AI Ultra subscribers.
Personalization draws from user history (with privacy controls) to tailor responses, such as dining suggestions based on past searches.
Users can share AI Mode sessions via links for collaborative planning, like trips, while maintaining control over shared content.
Targeting Key Markets: Languages & Demographics
The new languages target populous, high-growth regions, reaching over 1.2 billion speakers and potentially expanding AI Mode's user base by 15-20%.
In India, where English proficiency is low (around 10%), Hindi support could unlock billions of daily searches for education and e-commerce.
| Language | Primary Regions | Speakers (Native + Second Language) | Strategic Focus |
|---|---|---|---|
| Hindi | India, Nepal | ~609 million | E-commerce, education in India's 1.4B population. |
| Indonesian | Indonesia, Malaysia | ~200 million | Digital economy, local news in Southeast Asia. |
| Japanese | Japan | ~125 million | Tech, health for aging population. |
| Korean | South Korea | ~81 million | K-pop, finance for tech-savvy users. |
| Brazilian Portuguese | Brazil | ~250 million | E-commerce, events in Latin America's largest economy. |
Community reactions on platforms like X (formerly Twitter) show excitement, with users in Japan and Brazil praising cultural accuracy for local recommendations.
However, some early testers report rollout delays, with the feature still prompting "English only" in affected regions.
User Benefits & Broader Implications
Non-English users can now tackle complex tasks natively, such as comparing sleep trackers for elderly users in Tokyo (Japanese) or budget laptops with Hindi reviews in Mumbai.
Multimodal features enable voice queries or image uploads, like analyzing a skin condition for Korean health advice.
Privacy options include Incognito mode to prevent data use in training.
For businesses, AI Mode boosts conversions (4.4x higher value from AI visitors) but reduces overall traffic, raising concerns for publishers and SEO strategies.
Reddit communities discuss how it synthesizes answers with links but keeps users on Google longer, potentially eroding site visits.
Despite this, Google's approach emphasizes web discovery, with prominent links to sources.
Challenges: Accuracy, Traffic & Competition
While praised for speed and depth, AI Mode faces criticism for occasional inaccuracies, echoing early AI Overview issues like fabricated advice.
Reddit threads highlight user frustration with "hallucinations" in complex queries, though Gemini 2.5's benchmarks show improvements.
As it becomes the default experience, analysts predict shifts in SEO toward AI-optimized content.
FAQs
What is Google AI Mode?
AI Mode is an advanced search feature in Google that uses AI to provide conversational, nuanced answers to complex queries, integrating multimodal inputs like text, voice, and images, rather than just listing links.
Which new languages does AI Mode support?
As of September 8, 2025, AI Mode supports Hindi, Indonesian, Japanese, Korean, and Brazilian Portuguese, in addition to English, across over 180 countries.
How does Gemini 2.5 improve AI Mode for non-English users?
Gemini 2.5 enables local context understanding, handling cultural nuances and dialects without simple translations, while supporting advanced reasoning and multimodal features in the new languages.
Can AI Mode help with tasks like booking reservations?
Yes, agentic features allow bookings for restaurants and events, though currently limited to U.S. Google AI Ultra subscribers; global expansion is planned.
Will AI Mode reduce traffic to websites like Reddit?
Community discussions suggest it may decrease clicks by summarizing content directly, but it includes links to sources and could boost conversions for optimized sites.
Is AI Mode accurate in the new languages?
Some users praise its cultural accuracy, but some note occasional errors, and Google continues to improve via user feedback and model updates.
How do I access AI Mode in a new language?
Visit google.com/ai or use the Google app; it should detect your language preference, though some users report rollout delays—check for updates.
What are the privacy options in AI Mode?
Users can opt for Incognito mode to avoid data training, adjust personalization settings, and control shared links for collaborative sessions.
Does AI Mode support voice or image inputs in new languages?
Yes, multimodal capabilities like voice queries in Hindi or image analysis in Japanese are available, enhancing accessibility for diverse users.
What's next for AI Mode after this expansion?
Google plans more languages, default integration as the primary search experience, and expanded agentic tools like service appointments globally.
Follow Us on Facebook
