Spotting AI isn't getting any easier.

From American public media, this is Marketplace Tech.

I'm Lily Jammale. Miss and disinformation often get lumped together. They're not the same thing, but they are very much connected.

Say you hear that Christmas falls on December 23rd this year.

If someone told you that thinking it was true,

it's considered misinformation. but when it's spread with intent to deceive, that is

disinformation, which can easily get amplified by the folks in group one.

AI generated audio and video is everywhere this election season.

So before you click share, just know the tech used to create that stuff

is getting a lot better, a lot faster than you might think.

Says Joan Donovan, a long-time misinformation researcher and journalism professor at Boston University.

When it comes to the technology itself, what it shows is that we can realistically

depict any politician as having said or done something that never happened.

And we have that example now of the Robocall in New Hampshire that was already illegal to impersonate someone's voice, especially a politician, through

robo-calling, but now we've had to amend the regulation to also include impersonation by AI.

And is this just a lot more pervasive now because the tech has gotten so good?

I don't know. You know, one of the things that my research is starting to look at related to deep fakes is, is it the case that it's just a small number of actors producing a prolific number of deep fake video and deep fake audio.

You know, it's certainly gotten easier for people to sign up with different companies that offer AI impersonation as a service.

There's about a dozen of those companies now and it's increasingly harder and harder to know the difference and I think that as we

start to investigate and understand its effects on politics, we have to be very attuned to the fact that impersonation of political

candidates is illegal and can be

Disasterous even if it's just a prank

...