A Museum Uses AI to Save the Voices of the Warsaw Ghetto Uprising

The Publicis Groupe Poland team (Saatchi & Saatchi, Prodigious, MSL and Leo Burnett) created an artificial intelligence voice that captures all the emotions of a Holocaust witness, Stella Fidelseid, to tell her full story for the first time, in her own voice. Not only with her pitch, tone, accent, rhythm, articulation, and emotional expression, but also with real emotions such as pain, doubt, anger, fear, envy, sadness, despair, grief, fear, loneliness, trust, gratitude, love, and hope.

Stella Fidelseid, who died 20 years ago, was able to tell her memories for the first time ever, in her own voice and reflecting her emotion. Publicis Poland created a synthetic voice based on a recording of Stella from 1997 provided by the USC Shoah Foundation. And the team used Stella's voice to tell the world for the first time, her memories of the uprising, previously locked away in museum archives. AI tells it as Stella would have - with real emotions brought back to life.

The exhibition featuring the full story of Stella and other civilian witnesses of the Ghetto Uprising attracted more than 136 000 visitors and was the most popular exhibition in the history of the POLIN Museum of the History of Polish Jews. 80 years after the war, we were able to bring back to life the emotions of the Warsaw Ghetto Uprising, making them resonate for the next generations, continuing to remind us of the tragedy of the Holocaust.

And for Stella's granddaughter, the AI generated voice 'sounds exactly like her grandmother's'.