Fighting scams that depend on AI with AI.

From American Public Media, this is Marketplace Tech.

I'm Lily Jamali.

When they work, email phishing attacks can lead to big losses, adding up to $51 billion

over the past decade, according to the FBI.

And those losses are probably only going to get worse, because now a lot of those messages

are being written by AI.

Despite all the many positive things that artificial intelligence chatbots can do, they're

actually pretty good at writing up a convincing phishing email, too.

That could make it harder for targets to spot the red flags we've been told to watch

for on the news, or during those mandatory HR videos at work.

Now we, lowly humans, are enlisting an unlikely ally to help us.

Reporter Dina Temple-Raston from the Click Here podcast has been following one company

that's found a creative way to manage the problem.

AI is now capable of fixing the bad syntax and poor spelling that used to flag phishing

emails as fake, which has made it a lot more difficult for ordinary people to spot them.

Just ask Grant Walsh, he's the IT director at a company called Flow Control.

There was just an email came in to finance.

Hey, we had an updated making information, and that's financial person updated the wire

information and outgoing the chat for a six-figure number and the way that money goes.

But it wasn't just the money.

Walsh noticed that the company was getting emails that looked like they'd been written

...