Google is embedding its Gemini AI deeper into Google Workspace, headlined by a significant upgrade to Google Meet that introduces real-time speech translation designed to preserve the speaker’s authentic vocal qualities. Announced at the Google I/O 2025 conference on May 20, 2025, this feature initially supports English and Spanish, aiming to make multilingual virtual conversations feel more natural and inclusive. The company detailed that the AI-powered translation will begin rolling out in beta this week for Google AI Pro and Ultra subscribers.
This enhancement is part of a broader suite of Gemini-powered updates across Workspace. Gmail users can anticipate personalized smart replies later this year; this feature will analyze the context and tone of past emails and Drive files to generate responses that authentically reflect the user’s typical communication style. Additionally, Google Vids, the AI-assisted video creation tool, is now generally available to subscribers of Google AI Pro and the new Ultra plan. For users, these developments signal a shift towards more intuitive, intelligent tools capable of handling complex communication tasks, ultimately fostering improved productivity and collaboration. Google described its new translation technology as “very, very close” to enabling natural and free-flowing conversation.
Revolutionizing Cross-Language Collaboration
The new live translation in Google Meet moves beyond basic text-based captions, which was the previous method for in-meeting translation. The key innovation lies in the AI’s ability to match the speaker’s tone, cadence, and even expressions, thereby conveying intent more effectively. Dr. Evelyn Reed, a communications technology expert, said that “this isn’t just about words; it’s about conveying intent. Google’s focus on vocal nuance could be revolutionary for how distributed teams connect.”
To achieve this, Google trained its AI models on extensive multilingual speech datasets, encompassing diverse accents and speaking styles. Google say they focused heavily on minimizing delay to ensure conversations flow as naturally as possible, even with complex translations happening in the background, emphasizing the effort to maintain conversational flow.
While currently in beta for premium subscribers, Google plans an “early testing” phase for enterprise users later in 2025. This speech translation technology also underpins features in other advanced Google systems, such as the forthcoming Google Beam 3D video conferencing solution, which aims to create a “deeply immersive conversational experience.”
The AI Arms Race in Communication Platforms
Google’s latest enhancements arrive as competition intensifies among tech giants to integrate sophisticated AI into their communication platforms. Microsoft, for example, has been actively developing similar capabilities for its Teams platform. Its “Interpreter in Teams” feature, announced in late 2024, is designed to offer real-time speech-to-speech translation that can simulate the speaker’s voice for a more personal experience. Microsoft has also launched Translator Pro, a mobile application for enterprise-grade speech translation, and has been expanding live English subtitle translations across a wider array of Copilot+ PCs.
Meanwhile, DeepL, a prominent name in AI translation, introduced DeepL Voice in November 2024. DeepL’s approach, as its CEO Jarek Kutylowski explained at the time, focused on leveraging its renowned translation quality and security for real-time speech, initially prioritizing text-based output for captions to ensure minimal latency and solve immediate business challenges.
These parallel developments underscore a significant industry trend towards leveraging AI to dismantle language barriers, though strategies for voice preservation and output differ. For high-end systems like Google Beam, widespread adoption and cost justification will be a critical test.
Google’s Ongoing AI Integration in Workspace
The introduction of live speech translation with vocal nuance is the latest step in Google’s ongoing strategy to infuse its Workspace tools with AI. Previously, in August 2024, Google Meet had rolled out an AI-powered note-taking feature named “take notes for me” for certain Workspace users. This was part of an AI Meetings and Messaging add-on that also incorporated an auto-translation tool, representing an earlier iteration of translation assistance.
These features fall under the umbrella of Google’s Duet AI suite, which also includes functionalities like noise cancellation and video enhancement. While these AI tools offer significant benefits, the accuracy of AI-generated content remains an area of ongoing development and user attention. Further details on all the recent Gemini features across Google Workspace are available on the official Workspace blog.
