Respeecher Offers Celebrities Ukrainian Speaking Voice Clones for Video Messages Supporting Ukraine
Ukrainian synthetic speech startup Respeecher is asking celebrities to share messages of support for Ukraine in the Ukrainian language, but with their own recognizable voices. The startup’s voice cloning technology enables Respeecher to recreate someone’s voice and have that voice speak languages its originator may not know.
Respeecher is asking celebrities to let them produce a voice clone using new and existing audio recordings. The audio is used to train an AI model to mimic the original voice, even when saying things that were not recorded. After the celebrity approves the voice clone model, they will make a video sharing whatever message they want to send to those in Ukraine and supporters elsewhere. Respeecher isn’t offering any rules on the speech, just encouraging a sincere and personal message with concrete actions for other supporters to take. Respeecher then prepares a translation of the text in Ukrainian and a native speaker records the translated speech. The voice clone model overlays the new recording so it sounds as though the celebrity were speaking Ukrainian. The Ukrainian-speaking voice clone video (with English subtitles) will then be given to the celebrity for them to share online as both a statement of support and a way of pointing toward how people outside Ukraine can donate or otherwise help the war victims.
“Dozens of world celebrities and influencers used their voice to support Ukraine in this terrible war that Russia has started. The war against humanity and democracy. Respeecher wants to give these people a chance to use their voice in Ukrainian with the help of our state-of-the-art AI speech synthesis technology,” Respeecher wrote in a blog post about the project. Respeecher AI technology will help our Hollywood army of supporters to speak Ukrainian. With the same Emmy-award-winning voice cloning technology that was used in Mandalorian and many other big movies and TV shows.”
Respeecher does have Hollywood connections, as implied. The young Luke Skywalker’s recent appearance on The Mandalorian TV show spoke the words Respeecher plugged into its Mark Hamill voice clone. The de-aged face matched the de-aged voice, as the AI model was produced using recordings Hamill made decades ago. Respeecher also produced a deepfake Richard Nixon giving the speech he had prepared for if Apollo 11 had ended in tragedy. As the technology behind speech synthesis improves, these kinds of media uses are becoming more feasible and more popular. The translation aspect is starting to become more popular, too, as evidenced by efforts like the multi-language dubbing tool Aloud, made at Google’s in-house incubation hub, Area 120.