In 1902, when Frederick Gaisberg, sent by the Berliner Gram-O-Phone company, was wandering around Calcutta looking to capture Indian music for the newly-invented gramophone, he was taken to Gauhar Jaan’s concert. The legendary singer and tawaif known both for her captivating voice and charisma certainly left an impression. “In his memoirs, he recalls her as this well-dressed lady who, as all the Europeans are waiting, steps down and knowing how to charm people – puts her hand out and says ‘Good evening, gentlemen’. He’s quite stunned by her. Later, when he goes out for fresh air, he sees people perched on everything around – scaffolding, trees – to catch a glimpse of her, and realises her popularity,” says historian-author Vikram Sampath to the audience gathered at the launch of Nanna Hesaru Gauhar Jaan, the Kannada translation of his second book, My Name is Gauhar Jaan, the biography of a woman who would go on to be called India’s ‘gramophone girl’. What makes this translation unique though, is the fact that a majority of it was done through Sampath and techie Sandeep Singh’s AI translation technology, Naav AI.
Drawing parallels between the beginning of the AI age we find ourselves in and the time Gauhar Jaan was learning to navigate and poised to thrive in, Sampath continues, “When recording technology came to India, the male singers stayed away, and it was the women who overcame their inhibitions to record. Today, it’s a delicious irony that the first book translated through AI technology intervention is the biography of the first lady who used technology to record her voice.” After leading a lavish life, the songstress spent the last two years of her life in Mysuru, being paid `500 per month by the Maharaja of Mysore as a palace musician, before passing away in 1930.
The launch saw technologist Rohini Srivathsa and eminent musician Mysore Manjunath as guests of honour. Translator Shrigouri S Joshi who brought the human touch to the initial AI translation, gave an insight into the translation process, saying, “The draft translated through AI had 70 to 80 per cent accuracy.Many have asked what’s next since this technology is doing the work of translators and might replace us. This tool is to make our work easier. I’ve also worked on Vikram’s Waiting for Shiva; translating one chapter took three months. Understanding and translating legal language takes time. But with this tool, we finished the whole book in a month or two. Knowledge should flow freely from Kannada into other languages and other languages into Kannada.”
A demonstration of an in-progress audiobook version of Nanna Hesaru Gauhar Jaan, in which AI generated voices brought a section from the book alive, also wowed audiences.