BUFFALO, N.Y. — A recent report by the FBI says cybercrime losses hit nearly $21 billion last year. The FBI adds it’s a 26% jump year over year. The agency says AI is the primary reason.
“One way is that scammers are essentially combing through all of the available data about us, and there is a lot of data out there about us, and they’re using that information to create really targeted phishing attacks,” said Kimberly Palmer, a personal finance expert for NerdWallet.
She calls the rise in AI cyber scams disturbing.
“Basically, everyone is a potential target, and people’s identities are easy to find online,” Palmer said.
Palmer says the text and email scams are the first problem. The second is the growing trend of what’s known as “vishing.”
“Scammers are essentially using AI to create phone calls, audio and video calls that really look legitimate, like they’re from the FBI, the IRS, police,” Palmer warns.
So, take a beat the next time you get a suspicious call. Siwei Lyu, an AI expert at the University at Buffalo, was given a personal family photo to see just how easy a person’s information is to manipulate using AI.
“I think one of the biggest dangers is this increasing ability of generative AI models to create realistic, human voices, faces or any natural scenes [in] videos,” Lyu said.
Lyu says it can be generated under a minute, and scammers don’t need much.
“Ten seconds of your voice, one photograph or something, you know,” Lyu said. “Anything that this day we can get from a social platform about one person can be used potentially as training data for AI models for impersonation purposes.”
Using that family photo, Lyu was able to generate a video with a reaction to the Fenway Park sign and all.
“Once you have the data, you have a web page somewhere,” Lyu said. “You say, ‘clone this person’s voice and make that voice say these words.’ In a matter of I’ll say mostly a few minutes, sometimes even like half a minute, 30 seconds, you will get your sample back. They may not be perfect, but they are good enough to fool people.”
Lyu says AI has gotten so good that the typical tips of looking for eye blinking, a reaction or a reflection in the eyes won’t always work. He says now we need to tell ourselves that seeing is not believing.
“It’s time for putting more effort into verification,” he suggested.
He says it’s not a bad idea to have a secret phrase with family. Ask it if you get a suspicious call. Lyu predicts that those may turn to video chats in the near future.
“A real-time interaction, like in the Zoom call, you know, like a video conference,” he predicts.
Lyu says as AI continues to evolve, we should, too.
“As we get more information about this, having a better level, a higher level of awareness and better skills of verification, then we will have a lot more to protect ourselves,” Lyu said.
Palmer agrees.
“But I do think there is some hope because this is such a widespread problem, we could have some additional regulation and additional sophistication on behalf of people who are trying to crack down on this,” Palmer said.
Palmer says you can freeze your credit accounts at any time for free. This way, no one can open a new line of credit. You can unfreeze those at any time.
Palmer says your bank should be your first phone call if you fall victim. Then the police. The FBI and the Better Business Bureau are points of contact as well.
Click Here For The Original Source.
