Late last week, the Federal Communications Commission banned robocalls that use voices generated by artificial intelligence. This comes after phone calls during New Hampshire’s recent primary that used AI to imitate President Joe Biden and discourage voters from going out to the polls. Authorities have issued cease-and-desist orders against two Texas companies they say were involved with the robocalls.

To learn more, Marketplace’s Sabri Ben-Achour spoke with Wasim Khaled, CEO and co-founder of Blackbird.AI, a firm that uses good artificial intelligence to stop the bad. Below is an edited transcript of their conversation.

Sabri Ben-Achour: Let’s go back and hear that soundbite of the fake President Biden phone call. And just to be extra clear, this is fake, fake, fake, fake fake. This is not actually President Joe Biden speaking.

Ben-Achour: So that is not President Biden, but it is pretty convincing. This technology is pretty advanced, and it’s only getting better, right?

Wasim Khaled: Absolutely. The thing about today’s generative AI and technology infused landscape is the tools that are available to do the kind of thing you just heard here have become so low cost and so easy to use by threat actors, or anyone who’s looking to deceive the public, it’s made the ability to manipulate human perception so straightforward that it’s going to just continue to ramp up.

Ben-Achour: Are there less tangible or less obvious ways AI might be harnessed to manipulate people, whether that’s in election years or not?

Khaled: Well, it’s the combination. People fusing all of these things together to create much more convincing realities that can deceive people much easier. So you may note that very recently, there was a $25 million transaction that took place because the person who was making that wire transfer actually thought that their CEO and other executives were on a Zoom call and it turned out that the videos were superimposed over essentially avatars. They had simulated their voices using clips from their voices and it combined it all together to create something that was very, very difficult to overcome by the average person. And that’s because, today, people aren’t really ready for the kinds of mind-bending technologies that are out there.

Ben-Achour: That is nuts, a fake Zoom call — like fake video, fake audio, fake all of it? That’s wild.

Khaled: And it worked!

Ben-Achour: And it worked! Right. Do you have any thoughts on what we, as consumers should do to protect ourselves from this latest generation of malicious AI?

Khaled: When generative AI or any kind of media, you have to think that, yes, it very well could be fabricated, and that technology will get better and better. So video, audio — they’re all questionable. I would say the other thing is, for example, it’s happened to many people where they’ll get a distressed call from someone they know. And they say maybe they’re in Mexico and some foreign country, and they’ve been kidnapped, and they need money via a wire transfer or something of that nature. This is happening pretty widespread as well. This is an analog tip: If someone calls you and they say they’re in distress, they need money, or they need help or whatnot. Many people I know now have a codeword with their family. And so people really need to understand what’s happening in the information environment to protect themselves today.