By: Adrienne Viarengo
As technology has evolved so too have scams, becoming more sophisticated and thus targeting new subsets of the population in new ways. From mail scams, to cold calling phone scammers, to email phishing, scammers have always utilized burgeoning technology to try to hoodwink unsuspecting individuals.
Over the past few years, artificial intelligence has changed the landscape of many industries and introduced new security, privacy, and accuracy concerns. At one time, Disney’s storyline of Ursula stealing Ariel’s voice was far-fetched but now, due to the falling cost of the technology coupled with the proliferation of voices available from various social media platforms, criminal scamming ventures have begun successfully using deep fakes of voices.
Scammers use deep fake voice technology to trick unsuspecting consumers into trusting that a product or service is real by impersonating a popular celebrity or in other cases, someone the victim may know. Audio and video clips are spliced from real videos, manipulated, and reconfigured to fit a certain narrative. The result is an incredibly convincing portrayal of a target, either a celebrity or, in other cases, someone they may know, either asking for money or promoting a product, often without the target’s knowledge.
The Better Business Bureau warned of the increased prevalence of this type of scam in April of 2023, warning “[b]efore you make a purchase, take a minute to reexamine the post and social media account. . . . The photos and videos are most likely fake. If you make a purchase, you’ll lose money (often more than you expected) on a product that is substandard or doesn’t exist.”
In order to protect against deepfake scams, it’s important to slow down and take a closer look at the video promotion. Some telltale signs of an AI generated video include isolated blurry spots, odd edges around faces, changes in video quality, and unnatural blinking. Giveaways that audio is faked include unnatural phrasing, inflection, sentence structure, or background sounds. Additionally, significant pauses paired with faint keyboard clicks are an indication that a scammer may be typing the response to be fed into a voice-generating software.
Ultimately, the Better Business Bureau warns against making any consumer decisions based on viral videos and cautions that consumers should do their own research before purchasing something regardless of any perceived endorsement.
Taylor Swift, a celebrity globally known for her voice, was recently unwittingly used in a scam utilizing voice manipulation artificial intelligence to dupe her voice and overlay it onto a promotional video. Other celebrity likenesses recently exploited in this type of scam include Tom Hanks and Gayle King.
However, celebrities aren’t the only ones’ whose voices are getting cloned – the prevalence of social media use has resulted in scammers having unprecedented access to voice clips for many individuals. Artificial intelligence programs only need a 3-10 second sample in order to create realistic renditions of the target’s voice and can even layer that voice with emotions such as fear, urgency, sadness, or anger to further convince the target of its validity. One such scam is known as the “Grandparent Scam”. For example, an unsuspecting grandparent receives a panicked call from their grandchild, whose voice they recognize, and that grandchild describes an emergency which needs $500 wired right away. In that case, many grandparents might immediately comply and ask questions later. The scammer gets their money, the grandparents have experienced severe emotional distress, and their real grandchild is safe and unaware their voice was corrupted for this purpose.
So what can be done to protect against falling victim to such personalized and targeted scams? One possible precaution is to ensure that social media accounts are private and limit any public posts featuring your likeness or voice. Additionally, a good practice is to implement a sort of password or phrase to be used among close family members that can be invoked during an emergent conversation to ensure the person on the other end of the call is the person they seem to be. That way, if a family member is truly in an emergency, all parties can be assured that it is taken seriously and not a scam.
Regardless, it seems apparent that artificial intelligence will only continue to improve in its realistic abilities to clone imagery and voices, and with those improvements, the potential and prevalence for criminal scams will increase.
Student Bio: Adrienne Viarengo is a second-year J.D. candidate at Suffolk University Law School and is a staff writer for the Journal of High Technology Law. Adrienne has extensive experience working in Democratic politics and government, both in the U.S. Senate and on national and statewide campaigns. She received a Bachelor of Arts Degree in Political Science with a History minor and concentration in International Relations and a Bachelor of Arts Degree in English at the University of Massachusetts Amherst.
Disclaimer: The views expressed in this blog are the views of the author alone and do not represent the views of JHTL or Suffolk University Law School.
