Almost every couple of months, there is a news story about another voice cloning scam. It can come in the form of an urgent call claiming that a relative has been kidnapped. It could be about an impersonator calling on behalf of an organization’s boss to arrange a money transfer into their bank account. And sometimes, it could even be about an audio clip of a politician speaking in a way that they didn’t say.
Also read: I almost gave up on Google Gemini, but this one simple tweak made it useful
Voice technology firm Murf AI, which recently launched a TTS tool called Falcon, is trying to solve the problem in the simplest possible way by not offering voice cloning as a self-service product. Voice cloning is available only to enterprise customers who go through a sort of KYC (Know Your Customer) process. As CEO Ankur Edkie says, “it’s a capability that we have with our models, but we never open it up because of those risks.”
This list just keeps adding up. Each voice prompt has to contain an oral consent statement, identifying the individual and verifying that he or she is giving the approval for cloning. The recording has to be done live within the platform rather than uploading any recording because this makes it less likely that a pre-cloned voice could be used as the voice sample. The output is watermarking using the workspace ID and timestamp that remains even after compression and speed adjustments.
Also read: Before Resynced, every Assassin’s Creed game and the best order to play them
None of these steps is fail-safe, and Edkie does not make any claims about that. Murf uses an index of about a few thousand celebrities around the world as cross-reference when considering the prompts but admits that everyone can become a celebrity and hence the index cannot ever be complete. He compared it to how banks manage fraud risk, noting that “there’s always going to be some possibilities, but you have a way to kind of get hold of the situation.”
This is when things get awkward because of India. Edkie is clear that there is no voice rights legislation in India, as compared to certain places in America, wherein service providers, not just end-users, may face liability for allowing voice cloning fraud to occur. Murf has participated in regulatory consultations in the US but not in India. Edkie explained that “I don’t think there’s any explicit law that’s been created within India to solve these safeguards.”
Awareness about voice scams is present in Mumbai, Delhi, and Bengaluru, but not in other Indian cities in tier two or three. This may seem like an insignificant difference given the rapidity of growth in UPI and SMS scams in India.
There is another financial aspect associated with this process as well. The royalty-based licensing makes it so that Murf has a smaller selection of voices than its competitors. According to Edkie, this is an advantage, not a disadvantage, because “we are able to retain all the best ones, the ones that actually work for most customers.”
Whether that trade-off holds up at scale is the real test. Consent frameworks built by companies are useful, but they’re voluntary, and voluntary safeguards tend to bend under commercial pressure. The law hasn’t caught up yet. In India, it hasn’t even started.
Also read: Starlink vs Amazon Leo vs OneWeb: Who actually wins the Satellite Internet race in 2026