AI ‘voice cloning’ scams are on the rise. Here’s how to protect yourself

A mother from California recently fell victim to a scam that cost her thousands of dollars after receiving a call that sounded distressingly like her daughter. She now believes the call was the result of an artificial intelligence-generated hoax. Unfortunately, she is not alone; many individuals have faced similar threats from emerging 'voice cloning' scams, which leverage AI technology to create convincing replicas of anyone's voice from just a few seconds of audio. According to the Federal Bureau of Investigation, Americans lost over $893 million to various AI-related scams last year. These scams encompass not only voice cloning but also AI-generated phishing emails, romance scams, and other deceptive schemes. Scammers can impersonate anyone, including family members, friends, colleagues, or professionals, making it difficult for victims to discern the truth. Financial institutions, such as the UK's Starling Bank and the Commonwealth Bank of Australia, have issued warnings to their customers about the dangers of these scams. Experts highlight that AI voice replicas have become so realistic that distinguishing them from real human voices is increasingly challenging. "For the everyday person, it is just not fair to expect them to be able to spot this stuff," stated Henry Ajder, an AI-generated media consultant. "Even I struggle with it, and most people do." Scammers can craft an AI version of a person’s voice using a brief audio clip, often sourced from social media or previous scam calls. Social media also provides abundant information about potential targets, allowing fraudsters to create a convincing narrative. Commonly, these scams involve the impersonated loved one claiming to be in distress—such as being kidnapped or jailed—and urgently requesting money for their release. Gary Schildhorn, a Philadelphia lawyer who encountered an AI voice scam mimicking his son, recounted his experience: "There was no time to think. It was all, ‘I have to react to help my son. He’s in trouble.’" In more advanced scams, attackers may utilize text-to-speech tools or techniques like 'voice skinning' to manipulate their own voices in real time, enabling seamless conversations between the target and the AI-generated voice. Additionally, hackers may employ caller ID spoofing, making it appear that the call is coming from a familiar number, which adds another layer of deception. Previously, odd pauses or vocal inconsistencies were indicators of AI-generated voices. However, with technological advancements, these signs may no longer be present. Instead of focusing solely on the authenticity of the voice, experts recommend being vigilant for general scam indicators. Questions to consider include: Is there a sense of urgency? Is the caller insisting on secrecy? Are they requesting unusual methods for transferring large amounts of money? If you receive such suspicious calls, it's wise to contact your loved one through different channels—like texting, using another phone, or reaching out to mutual acquaintances. Deborah Del Mastro, the California mother who was scammed, only reached out to her daughter after sending money to the fraudsters. Fortunately, her daughter was at work and unharmed. Establishing a precautionary 'code word' among family members or coworkers can help verify identities and prevent scams. This word or phrase should be known only to a small group of people and should not be easily found online. As Ajder advises, "If you suspect something may be off, it’s far better to have a loved one laugh at your concern than to risk running to an ATM."