Music streaming platform Spotify is testing an AI-powered Voice Translation tool for podcasts. The tool can translate podcasts into different languages using the podcaster’s own voice, creating a more authentic listening experience. Spotify has collaborated with popular podcasters such as Dax Shepard and Lex Fridman to generate translations in languages like Spanish, French, and German. The platform plans to expand the tool to include other shows and launch translated episodes worldwide for both Premium and Free users.
The new tool harnesses the latest advancements in AI, including voice generation technology from OpenAI. OpenAI’s ChatGPT has developed new voice and image capabilities that enable it to see, hear, and speak. The tool utilizes custom text-to-speech models created in collaboration with professional voice actors. Spotify’s pilot voice-translated episodes will first be released in Spanish, followed by French and German. The company sees this as just the start of its commitment to using AI to empower creators and expand their storytelling capabilities.
Spotify’s Voice Translation tool for podcasts follows the platform’s previous endeavors into AI-powered audio. In May, they introduced an AI ‘DJ’ feature that delivers personalized playlists accompanied by commentaries narrated in a realistic AI voice. This feature was made possible through Spotify’s acquisition of AI voice startup Sonantic. These advancements raise the question of when similar technology will be available for artists to release tracks in multiple languages, maintaining the artist’s own voice. South Korean entertainment giant HYBE showcased this potential earlier this year with a multilingual track produced using fake voice AI technology from Supertone, a company HYBE acquired.