A recent report highlighted the prevalence of artificial intelligence (AI) backed voice scams, which seem to be far more in numbers in India as compared to other major countries such as the US, UK, Germany and France.
About 47 per cent of Indian adults surveyed report having experienced or knowing someone who has experienced some kind of AI voice scam. This is almost double the global average at 25 per cent, mentioned the McAfee report. The victims of AI voice scams (83 per cent) said they lost some amount of money and about 48 per cent of them reported a loss of Rs 50,000 and above.
“Artificial Intelligence brings incredible opportunities, but with any technology there is always the potential for it to be used maliciously in the wrong hands. This is what we’re seeing today with the access and ease of use of AI tools helping cybercriminals to scale their efforts in increasingly convincing ways,” said Steve Grobman, McAfee CTO.
AI Voice Cloning
With as many as 86 per cent of Indian adults sharing their voice data online or in recorded notes, at least once a week (via social media, voice notes and more), cloning someone’s voice is now a powerful tool in the arsenal of a cybercriminal, mentioned the report.
It also revealed that scammers are using AI technology to clone voices and then send fake voicemails or voice notes or place calls directly to the victim’s contacts pretending to be in distress. With 69 per cent of Indian adults not confident in identifying the cloned version of their voice, it is no surprise that this technique is gaining momentum in the country.
“Should you receive a call from your spouse or a family member in distress and asking for money, verify the caller – use a codeword, or ask a question only they would know. Identity and privacy protection services will also help limit the digital footprint of personal information that a criminal can use to develop a compelling narrative when creating a voice clone,” advised Grobman.
McAfee researchers found more than a dozen freely available AI voice-cloning tools on the internet. One of the tools (paid/unpaid) found by researchers could produce a voice clone (85 per cent match) using just three seconds of audio.
By training the data models, the researchers were also able to achieve a 95 per cent voice match based on just a small number of video files.
Using the cloning tools they found, researchers discovered that they had no trouble replicating accents from around the world, whether they were from the US, UK, India or Australia. But more distinctive voices were more challenging to copy.