At its annual Google I/O 2025 event, Google unveiled one of its most exciting updates yet for global communication: real-time speech translation in Google Meet. Now available in beta for Google One AI Pro subscribers, this feature enables natural, free-flowing multilingual conversations during video calls—with voices, tones, and expressions preserved in the translated audio.
This breakthrough leverages an advanced audio language model developed by Google DeepMind, which translates spoken words into the listener’s chosen language as they’re being said. The translated speech plays over the original voice, which is still faintly audible to maintain context and tone. Imagine hearing someone speak Spanish while their translated English voice, still capturing their intonation and emotion, is layered seamlessly on top.
“Think of it like having a live interpreter in the room—but instead of someone else’s voice, it’s yours,” said Yulie Kwon Kim, VP of Product at Google Workspace.
The tool is designed to make multilingual conversations smoother and more accessible in real time. Currently, it supports English and Spanish, with additional languages like Italian, German, and Portuguese expected in the coming weeks. Google emphasized that this feature is especially useful for personal and professional communication alike—from grandchildren chatting with non-English-speaking grandparents to cross-border teams collaborating without language barriers.
How It Works
Once a Google AI Pro subscriber enables the feature, everyone in the meeting benefits from real-time translation. Users simply activate the “Speech Translation” option in Google Meet, and the system does the rest—transforming the speaker’s words into the target language with minimal latency.
Google says that this innovation builds upon its longstanding work in Google Translate but takes it to a new level by integrating live voice with emotional cues, cadence, and personality. For example, a business leader’s passion during a product pitch, or a family member’s excitement about an upcoming trip, comes through clearly even in translation.
Who Can Access It—and What’s Next
The beta version of this feature is now available to Google One AI Pro (formerly AI Premium) subscribers for $20/month. Enterprise users with Workspace accounts can expect early testing opportunities later this year. The rollout marks a pivotal shift in how Google envisions communication in a world where language is no longer a barrier.
“We’re enabling meaningful conversations that would have previously required a human translator—or simply never happened at all,” the company said in a blog post.
A Future Without Language Barriers?
While the promise of AI voice translation is exciting, it raises thought-provoking questions. Will future generations lose interest in learning new languages if AI can translate for them? Perhaps. But Google is positioning this not as a replacement for language learning, but as a bridge to connection—a way to foster understanding instantly, when and where it’s needed most.
As this technology continues to evolve, Google is actively seeking user feedback through its product research programs and invites users to rate their experience in real time during calls.
For now, the ability to converse across languages—hearing yourself, only translated—is no longer science fiction. It’s Google Meet in 2025.
