Rising Bharat Summit 2025: ElevenLabs walks tightrope between AI voice cloning innovation, privacy concerns

Siddharth Srinivasan, go-to-market leader for India at ElevenLabs.

Amid growing concerns around deepfakes and misuse of artificial intelligence (AI)-generated voices, AI voice synthesis company ElevenLabs is doubling down on building guardrails into its voice cloning technology.

The two-year-old London-headquartered startup, which recently made headlines for real-time voice dubbing of Prime Minister Narendra Modi on the Lex Fridman podcast, says the firm is absolutely aware of the privacy risks that come with cloning a person’s voice.

“From day one, we realised that responsibility has to be part of the framework,” said Siddharth Srinivasan, go-to-market leader for India at ElevenLabs while speaking at the Rising Bharat Summit 2025 on April 9 in New Delhi. ElevenLabs has built strict user consent protocols, digital watermarking of all generated content, and back-tracking systems that can identify the source and purpose of any voice clone.

“The average person can't go up onto ElevenLabs platform and start generating clones, creating mischief with that. We've got processes where first of all, you yourself are in the room and consenting. Number two, every generation is watermarked. And third, if there is a need to track back a piece of content and verify whether it was done, and what was the source of generation, we make that happen the full way. And that's a 99% precision,” he added.

These guardrails are especially crucial as the startup scales. In the Modi-Lex Fridman podcast, ElevenLabs cloned Modi’s voice with his consent to deliver a three-hour conversation same day in English, using his own voice model.

Moreover, ElevenLabs is betting that clear use cases and transparency will win public trust. He argued that cloned voices, when controlled by the IP owner, could be “liberating,” implying creators can scale content across languages and markets.

“It's a liberator for you. You have your voice clone, and then you get to choose what the purpose is, and you get to use it in a way which makes sense for you and amplifies everything that you're doing,” he further added.

This balance between power and protection is also echoed in the broader AI ecosystem.

Nothing, the smartphone startup co-founded by ex-OnePlus executive Akis Evangelidis, is integrating AI to personalise the smartphone experience while also acknowledging data privacy concerns.

“The user hasn’t yet been educated or seen clear use cases which benefit them,” Evangelidis said. “But when experiences are personalised, contextual, and value-adding, people will gradually worry less about privacy.”

Also, read: Inside the tech behind Modi-Lex podcast: ElevenLabs India head explains the breakneck translation speed