DeepL Expands Real-Time Voice Translation to Meetings and Events

According to AI TIMES, DeepL said on September 16 that it unveiled an update to its real-time voice translation service that preserves the speaker's voice, emotion, and tone, while supporting voice-to-voice translation in more than 30 languages. The desktop app released alongside it connects with major video conferencing platforms, allowing users to use translated subtitles or voice translation immediately in multilingual meetings without separate, complicated setup. At the core of the change is an effort to go beyond translating the meaning of words and also convey how and by whom they are spoken. It is also designed to distinguish and maintain speaker-specific characteristics in meetings with multiple participants, and it signals a push to expand use cases from enterprise meetings to large-scale on-site events. The move is seen as a sign that the basis of technology competition is shifting from accuracy and speed to how well context and expression in conversation can be preserved. At the same time, as features that reproduce vocal characteristics become more advanced, challenges around consent, data use, and trust are likely to grow as well.