Bloomberg Taps Video Synthetic Speech AI Startup Papercup to Dub News into Spanish
Bloomberg Media has teamed with synthetic speech and translation startup Papercup to produce Spanish-language versions of its YouTube videos. Papercup will redub the videos with AI-generated Spanish voices as a way of localizing hundreds of hours of content for Spanish-speaking viewers in Latin America and the United States.
Vaso de Papel
“At Papercup, our mission is to make the world’s videos watchable in any language. There are billions of hours of video stuck in a single language – our AI dubbing tool will unlock vast volumes of this content for a global audience,” Papercup CEO Jesse Shemen said. “I’m excited about working with Bloomberg given their unparalleled track record of producing financial news content that an entire industry depends on. I have no doubt that audiences across the globe want to engage with and share their content – I’m looking forward to Papercup playing a part of that journey.”
The London-based Papercup uses AI and natural language processing to transcribe speech into text, then translates it into another language and synthesizes a version of the original speaker’s voice to recite the script in its new tongue. Bloomberg’s many videos are undergoing the process, resulting in news reports, financial analyses, and documentaries that sound like the original speakers learned Spanish and rerecorded the videos in their new language. The news organization sees Papercup’s technology as a way to vastly expand its potential audience to millions of new viewers without investing nearly as much time and money as is usually necessary to localize content.
“As a global news leader, it’s Bloomberg’s mission to ensure that reliable, trustworthy news can reach as many people as possible,” Bloomberg Media video and audio general manager Travis Winkler said. “Papercup’s advanced AI tools give us an exciting new way to reach global audiences, and we expect to see audiences respond enthusiastically.”
Papercup’s approach to AI translation and dubbing is becoming more popular. Even with human auditors to check the accuracy, it’s far cheaper and more efficient than manual transcription and translation, plus the synthetic voices lend authenticity by sounding like the original speakers. Discovery and Sky News are already Papercup clients. Projects like Google’s Aloud multi-language dubbing tool and Veritone’s deal to reproduce some iHeartMedia podcasts in different languages demonstrate the market for it, as does Respeecher, whose synthetic voices are shaping content as wide-ranging as BBC Radio and Star Wars. Papercup raised $20 million back in June but has a lot of opportunity ahead of it.
“Synthetic media sometimes is used to create entirely new content, and other times it is simply modifying existing content into a new format,” Voicebot founder Bret Kinsella pointed out in our Synthedia newsletter. “Localization falls into the latter category and will be more and more attractive to companies that want to extract more value from each asset they produce, along with harvesting new value from their catalog of spoken language content.”