The on-device feature aims to remove language barriers, though experts flag accuracy concerns.

A major technology company unveiled a new smartphone feature on Monday that translates spoken conversations between two languages almost instantly, without sending audio to an external server.

The system uses an on-device AI model to interpret and translate speech within roughly a second, and the company says it currently supports twelve languages, with more planned for a future update.

Industry analysts see the feature as part of a broader push toward on-device AI, which processes data locally rather than in the cloud, offering better privacy and working without an internet connection.

However, some linguists caution that the tool still struggles with idioms, regional accents and fast-paced speech, and they recommend it as a supplement to, not a replacement for, human interpreters in formal settings.