Picture this scenario: The escrow agent receives a call from you, his trusted lender. It sounds exactly like you — so much so that he’d swear on a stack of Bibles it was you.
The "lender" asks for some urgent changes in the wire instructions for an upcoming closing, following up the request with an email verifying the details outlined in the phone call. The email, by the way, looks perfect.
What does the escrow officer do? After all, everything looks and sounds legit.
If he acts on this all-too-real request, you — the actual lender — and he have just been had. Because it wasn’t you who called or sent the email.
Fraud using artificial intelligence could reach a massive $40 billion by 2027, warns a new white paper.
So if the escrow agent doesn’t ask for a face-to-face verification and take other security measures, you’ve both become a victim of a deepfake, a growing type of mortgage fraud that uses fabricated video, audio, and images that are both “incredibly real” and hard to discern from the real thing, according to a new white paper.
Actually, the scenario described above was AI-generated. And the criminals got away with nearly three-quarters of a million dollars.
Fraud using artificial intelligence could reach $40 billion by 2027, warns the report, “How AI Is Changing Fraud in Real Estate." And it’s relatively easy to pull off, says the report by Closinglock, a fraud prevention platform used primarily by title companies.
“Cyber criminals can readily pull your voice from podcasts, social media, YouTube, and other easy-to-access channels,” the paper reads. “Just three seconds of audio can create an 85% voice match... Criminals can generate your face, words, voice, and moving image.”
“AI will continue to advance and evolve... The people and systems involved in real estate transactions must get more savvy about AI and fraud prevention."
—Closinglock white paper
While some attempts at deepfakes are thwarted, others have duped many companies and real estate professionals. Fraudsters are now constantly targeting title companies, realty firms, and individuals, according to the report. Last year alone, it says, the cybercrime price tag hit a staggering $9.5 trillion in all its iterations.
Real estate is particularly vulnerable for several reasons:
- There’s always a sense of urgency in real estate transactions.
- The people involved are often very emotional.
- The transaction values are high.
- Many transactions and documents are public record.
Put all that together, the report notes, and you have “the perfect conditions for deepfakes.”
Another key factor for the ruse’s success: AI is evolving so fast, the white paper notes, that many title companies “are scrambling to catch up.”
But businesses can protect themselves. Among other things, they must rethink the way they manage identity verifications, codify their communication protocols, and ask vendors about their AI-fraud readiness.
And if you plan to bring a new tech vendor on board, you need to be particularly careful.
Several questions to ask before moving forward, according to the report:
- Do you verify beyond static image?
- How do you detect deepfakes or spoofed voice calls?
- How do you use AI to prevent fraud?
- What do you do if you suspect fraud?
“AI will continue to advance and evolve,” the paper emphasizes. “In turn, the people and systems involved in real estate transactions must get more savvy about AI and fraud prevention. Building trust with these same people and systems can support both title professionals and the industry at large for decades to come.”
