Receiving a random call from a family member, passionately explaining they’ve been in a horrific car accident and urgently need you to send money or they’ll go to jail, can be distressing. Even though the voice sounds convincing and the call comes from their number, something feels off.
You decide to hang up and call back, only to surprise them with your call, as they haven’t left their house and haven’t been in any accident. What you experienced was an AI-powered scam call. Though the threat of AI-enabled scams can be terrifying, there are steps you can take to protect yourself and your loved ones.
Hardest to Detect
Last March, The Washington Post highlighted scammers using AI models to mimic the voices of relatives and friends, asking for help, then defrauding them of thousands of dollars. Some AI software needs only a few sentences to produce convincing dialogue, capturing the tone and emotional nuances of the speaker, while others require just three seconds of their voice. For older victims, it’s often challenging to discern if the voice is real or synthetic, even when the emergency described by the scammer seems implausible.
With advancements in AI text-to-speech technology, it’s easier and cheaper for scammers to produce fake yet convincing voices, with the latest models capable of speaking multiple languages. OpenAI recently announced a new model for text-to-speech conversion, broadening the scope of voice cloning.
Different Approaches
Here, you’ll need to think calmly and employ other methods to detect the scam, such as hanging up and calling back. Scammers rely on social engineering tactics, influencing and pressuring individuals into making quick decisions. Therefore, maintaining calm and employing logical thinking is crucial.
One well-known security tip, suggested by experts, is establishing a secret safety word with close family members. You could also ask a personal question only the real person would know, like what they had for dinner last night or what they were wearing when they left the house.
Zero Trust Security Model
This approach means not trusting anything or anyone blindly and verifying the reliability of every device, user, service, or anything else before granting access to your data. It involves periodically re-evaluating that trust to ensure nothing has been compromised during use.
This model can be applied to such phone calls. Always verify the source of any information reaching you and think logically about the situation. Since scammers rely on victims not thinking critically, a little logical thinking will often foil their plans.
Leave a Reply