Google continues to push the boundaries of artificial intelligence, and its latest announcements at I/O 2025 put Gemini in the spotlight with innovations that radically transform our daily communications.

Among the stars of the show: real-time translation for phone and video conversations, capable not only of instantly converting words but also of cloning your interlocutor’s voice for an ultra-natural experience.

Let’s take a concrete example: imagine calling a business partner in China who only speaks Mandarin. Thanks to the new Voice Translate feature on the Pixel 10, Gemini analyzes the voice in real time, translates the speech into English (or one of the dozens of supported languages), and synthesizes it all with the timbre, tone, and even the original emotions of your correspondent.

Result? You hear their familiar voice expressing itself in the language of Shakespeare, with no perceptible latency, as if the language barrier never existed.

This tech, powered by Gemini’s advanced models, goes beyond simple textual translation: it preserves intonation to avoid cultural misunderstandings and makes exchanges more human.

And it’s not limited to phone calls. Google is extending this magic to Meet for video conferences, where the AI translates in real time starting with English and Spanish, with plans for rapid expansion to other languages like French, Chinese, or Hindi.

Beyond the wow factor, this advancement could revolutionize international business, travel, and even personal relationships by making the world more connected.

But beware: ethical questions are arising about the privacy of cloned voices. Google assures secure data handling, but the future will tell if this mind-blowing AI doesn’t cross any red lines.