“Welcome to Vistara. This is a pre-boarding announcement for passengers flying to Delhi via flight 6130.”
These ubiquitous voices are often overlooked, but the individuals behind them, voice-over artists, constitute a booming industry in India with over 20,000 artists and 3,000 production studios in India.
But AI voice cloning is now deeply hurting this profession and the artists. A GenAI model trained over 10-15 sentences in a person’s voice can create a 5-hour long audiobook within minutes or even dub an entire OTT web series in 10 different languages. Voice over artists are getting random calls to «steal their voices».
The Association of Voice Artists (AVA) in India is now campaigning against the malpractices which are using the audition recordings, telephone conversations or even publicly available audio of voice-over artists to create their AI-generated voice clones.
Also read | AI.con: a new voice for Jamtara-type scammers
“Unfortunately, there are many free online tools which can create voice clones of known artists and there is no legal recourse to such infringement,” said Darrpan Mehta, a veteran voice artist and founder of Sugar Mediaz and India Voice Fest.
The AVA has reached out to industry body FICCI to represent its concerns before the IT Ministry to formulate policies to prevent voice deepfaking for commercial exploitation.
“As an industry, we are urging the government to recognise that similar to videos and