News

Scammers use AI to clone kids’ voices

b5aa8c1a-f512-47f1-9319-ec028c5d54a7

When a father in India received a call from an unknown overseas number last month, he did not expect to be the latest to fall for an elaborate fraud scheme involving Artificial Intelligence (AI).

The scammer, who claimed he was a police officer, told Mr Himanshu Shekhar Singh that his 18-year-old son had been caught with a gang of rapists and needed Rs30,000 ($486) so he could clear his name, Indian media reported on Monday.

Mr Singh told The Indian Express of the Jan 8 incident: “The next minute, I heard a voice saying, ‘Papa please pay him, they are real policemen, please save me’.

“I could not doubt for a second that he was not my boy. The style of speaking, crying – everything was the same.”

While Mr Singh was suspicious, he feared the caller was a kidnapper, so he made an initial payment of Rs10,000.

He then decided to look for his son on his own. The teenager was later found taking a test at an education centre unharmed.

This was one of three prominent cases to rock New Delhi and its National Capital Region in recent weeks as scammers tap into AI to develop convincing fake voice recordings of children to trick their parents into transferring money.

Audio deepfakes can be created with a real clip of the target’s voice.

In another case, Mr Rajesh Kumar Garg, an engineer who lives in Pitampura, received a call on Jan 9 from a person threatening to implicate his son, who is pursuing his MBA in Hyderabad, in a rape case.

“They said my son was caught with a gang involved in heinous crimes… but as he seemed to be from a good family, they intended to release him – provided I pay them,” Mr Garg told The Indian Express.

“I heard my son crying over the phone, it sounded exactly like his voice. First, they demanded Rs50,000. When I said I didn’t have that kind of money, they asked me to pay Rs30,000, and I did as they asked.”

A friend then suggested that Mr Garg call his son.

“My son said he was in college and was fine. It was a +92 code with a police officer’s photo,” said Mr Garg.

In a similar case, a journalist living in Noida, Uttar Pradesh, received a scam call where the scammers used technology to mimic her son’s voice.

Fortunately, her son was studying in front of her when the scammers called.

The journalist said: “It is a huge concern that cyber criminals are targeting children now. Where are they getting details of kids and their parents? This must be thoroughly investigated with utmost priority.”

Delhi police officer Manish Kumar Mishra said: “Such cases are not very frequent, but recently there has been an uptick in cases of ‘cloning’. We are trying to understand how exactly cyber criminals are creating cloned voices to dupe people.”

With rapidly improving AI, voice cloning has become more convincing. “It takes as little as three seconds for your voice to be duplicated,” said chief data officer of Transaction Network Services Greg Bohl.

“When you start saying, ‘Here’s my address,’ or ‘I didn’t make that appointment’, when you start getting sentences going, they have you.”

He said voice-cloning capabilities are rapidly improving, including the ability to type in real-time text strings for a cloned voice to repeat (enabling live-spoofing during calls) and simulating noisy environments.

“They can bring in the sound of a subway station, an airport,” he said. “You’re at the airport; you’re the executive travelling. You’ve been in a car accident, and you’ve got that street noise going on behind you. It sounds exactly like what’s taking place.”

Another concerning twist? “The scammers have the ability to add accents and age ranges,” he said.

McAfee’s recent report called The Artificial Imposter highlighted that a majority of Indians (69 per cent) cannot distinguish between an AI voice and a real voice.

Nearly half of Indian adults have either experienced or know someone who has experienced some sort of AI voice scam, which is almost twice the global average of 25 per cent.

Some 83 per cent of Indian victims reported a monetary loss, with 48 per cent losing more than Rs50,000.

Indo-Asian News Service

quote-icom
“It takes as little as three seconds for your voice to be duplicated. When you start getting sentences going, they have you.”
Chief data officer of Transaction Network Services Greg Bohl
promote-epaper-desk
Read this week’s digital edition of Tabla! online
Read our ePaper