How AI Is Supercharging Investment Fraud (And What You Can Do About It)

Marguerita Biel (0:00)

Artificial intelligence is giving fraudsters powerful new tools. Deep fake videos that can put words in anyone's mouth. Voice cloning that can mimic a loved one from just a few seconds of audio. Synthetic identities built to pass every verification check you threw at them. The technology is moving fast, but the underlying playbook hasn't changed. Urgency, emotional manipulation, pressure to act before you think. Fraudsters have always relied on those levers. AI just makes them easier to deploy, harder to detect, and far more convincing.

On this episode, we'll explore how AI has been reshaping the fraud landscape for investors and member firms, what FINRA is seeing from its unique position at the center of the markets, and what FINRA is actively doing on the investigative side, the intelligence side, and the investor education side, to stay ahead of these threats and protect investors and member firms.

Welcome to FINRA Unscripted, I'm your host, Marguerita Biel. Joining us today to talk about AI-enabled fraud are Christine Kieffer, Senior Vice President of Investor Education and President of the FINRA Investor Education Foundation, and Mayur Patel, Director, Financial Crimes Intelligence, Strategic Intelligence. Mayur and Christine, welcome to the podcast.

Christine Kieffer (1:28)

Thank you.

Mayur Patel (1:29)

Hi, Marguerita. Happy to be here.

Marguerita Biel (1:32)

Great.

So let's dive right in. Is AI-enabled fraud really a new problem, or is it a new version of something we've been fighting for a long time?

Christine Kieffer (1:43)

Well, I'll start. The short answer is it's both, and that distinction matters in how we fight it.

I'll frame it this way. The what is very familiar, but the how has fundamentally changed. So the types of schemes that we see, the fraud schemes, that's the what. They haven't really changed. We still see variations of pump and dump schemes, pitches involving non-existent investments, and those that capitalize on today's trending topics to otherwise defraud investors. But how fraudsters are executing those schemes has completely changed with the use of AI.

We're well past the days of e-mails with typos and videos with obvious inconsistencies and facial movements that don't match words or six fingers on a hand. Call scripts are perfected. Voices are changed and impersonated. Whole websites are built within minutes. Legitimate people and businesses are being cloned. And this builds source credibility for the scammer. It's simply a shortcut to trust.

So many fraud schemes FINRA sees rely on fraudsters needing to build this trust and to borrow trust that belongs to someone else. And AI has simply lowered the cost and skills to do it convincingly and at scale and lightning speed.

Marguerita Biel (3:03)

Great. Thanks, Christine.

So walk us through the main categories of AI-enabled fraud that you're seeing. I know there's deepfakes, voice cloning, social media impersonation, synthetic indebtedness, and that's just to name a few. What are the most prevalent right now?

Mayur Patel (3:21)

Yeah. So Christine started off great describing how she frames it, the what and the how.

The what, which she talked about, is familiar. All the different types of scams that FINRA and we as a society have been seeing over the years and decades, whether that's imposter related scams, romance, investment club, fake websites, mobile applications that are fake. Then of course, the cyber enabled fraud scams like account takeovers, business email compromise. That's the what. These are familiar with many investors. They're familiar with the industry and in general, like I said, our society. But the how has changed and of course, AI is supercharging all of this, especially in the context of let's say social engineering, for instance. These are tricks that fraudsters use to manipulate people into giving away something, whether that's private information, money or access, access to their brokerage accounts or access to accounts in general. AI is being used to supercharge all of this. And so what's most prevalent now from a FINRA perspective, it's how AI is being used to make fake ID documents, for instance, to help in the impersonation scams or deep fakes of both images and voice calls. And I think the social engineering aspect is very key into three specific things that I want to talk about. And I'll define a little because many people might not be aware of them, but phishing, smishing, and vishing. These terms are thrown around, but when it comes to phishing, these are emails and websites that are sent directly to individuals, to in essence, steal information. AI is being used to create and supercharge these. Smishing is just text messages that I think all of us receive.

We've seen indications where AI is being used to craft the language and send it out to mass people. And then finally, phishing, these are voice clones that we talked about that can be deep faked very, very easily. That's what we are seeing that is most prevalent right now in the market.