Hollywood screenwriters are currently striking en masse, protesting, among other things, the use of artificial intelligence (AI) technologies to produce movie and television scripts, potentially putting human writers out of business.
Worse, the same technologies that are churning out copy, such as ChatGPT, or creating illustrations, like Dall-E, are enabling cyber-criminals to prey on unsuspecting consumers, using AI tools to replicate voices and images that fool people into thinking they are communicating with someone they know when they are actually speaking to a scammer.
One scam that has been around for a while, for example, a phone call allegedly from a child or grandchild pleading for money has been elevated to an even more frightening level as fraudsters use online images and audio clips of the real person to convince a victim to respond. The technology presents a threat to businesses as well, as employees may be unable to distinguish the cloned voice of a boss or colleague from a deep fake scammer.
Last year alone, more than $10 billion in losses from online scams were reported to the FBI. AI researchers are developing software tools, such as D-CAPTCHA, designed to recognize AI-generated voices and images in an effort to fend off cybercriminals. But these systems have not yet been commercialized.
In the meantime, consumers need to look and listen more critically than ever at communications that could be suspect,such as a call allegedly from someone you know asking for money or personal information.
According to Los Angeles Times reporter Jon Healey, another way to foil potential scammers is offered by Haywood Talcove, chief executive of the LexisNexis Risk Solutions Government Group. Talcove suggested choosing a unique but easily memorized code word—one that is unlikely to be guessed by scammers—and giving it to everyone in the family. Then, if a call comes in from someone claiming to be a relative, ask for the code word before responding. Even if a suspect caller does not claim to be a relative, the very fact that you ask for a code word may be enough for the thief to end the call.
