October 2026

Confronting the Age of AI Deepfakes

Deepfakes are flooding the internet, making common scams harder to spot, and making cyber insurance a critical tool for managing fallout.

This blog helps you answer the following questions:

How do AI deepfakes change what “urgent wire” (and similar) scams look like, and why are they harder to spot now?
What repeatable verification steps should staff follow before approving wires, vendor changes, or other high-risk requests in a “deepfake world”?
What should you do immediately when you suspect an AI deepfake to prevent loss, preserve evidence, and escalate appropriately?
How could cyber insurance help after an AI deepfake or social engineering incident, and what policy details matter most?
What are the benefits of cyber insurance?

Imagine this scenario. Maya, a family’s personal assistant, is in the U.S. when she receives an urgent FaceTime call and voicemail from her boss, Mr. James Smith, who is traveling abroad. The call and voicemail authorize a wire for an art purchase and give an account number ending in 1234.

The FaceTime call looked convincing, but something felt off to Maya, so she paused and called Mr. Smith’s office number. Mr. Smith answered. He explained that he never made the request.

After looking into the matter, the clips used in the call were confirmed as AI-synthesized video stitched together from public footage of Mr. Smith, who is a well-known public figure.

This scenario is real, and while Maya acted quickly and confirmed with Mr. Smith before taking actions that could have resulted in financial loss, it reflects a broader shift in the sophistication of cyber scams. Gone are the days of poorly written emails from strangers; increasingly these scams now present a familiar face and voice, along with plausible details that even discerning individuals may fall for— and that raises the risk level.

Explore how AI-powered deepfakes are changing the cyber world, creating new security risks, and what you can do to defend against them.

Why AI Deepfakes Raise the Stakes

Generative AI, or Gen AI, reduces the time and skill required to produce believable fake audio and video. This type of AI creates original content like text, images, audio, and video using deep learning models that attempt to mirror the human brain’s decision-making process. While this technology is fascinating and has many legitimate use cases, cyber attackers are using it to create realistic images, videos, and audio (or clone them) to pressure people into authorizing actions, such as making a bank transfer or sharing personal information like a Social Security number.

The Federal Trade Commission (FTC) issued a challenge in 2023 for good actors to find solutions to the rise in voice-cloning scams. In the years since, technology has only gotten faster and more realistic, compounding the problem. “In 2025, people reported losing about $16 billion to scams, compared to $12.8 billion the previous year,” according to the FTC. And unsurprisingly, the top reported type of scam is imposter scams, where someone pretends, whether through AI or otherwise, to be a known entity.

For high-net-worth individuals, households, family offices, and people in high-visibility roles like athletes or CEOs, the risk rises. Their day-to-day lifestyle often involves high-value transactions, delegating authority, or appearing in public. Yet for the cyberattacker, the goal remains the same: one successful approval or transaction—only now they are emboldened by technology that makes it quick and easy to sound and look like a legitimate source.

Create Your Own Verification Workflow to Find Deepfakes and Scams

In our scenario, Maya’s response worked because she used a separate, trusted channel to confirm the request before acting. That verification stops many impersonation attempts and is one to add to your own toolbox.

Family members, staff, and anyone involved in day-to-day operations and transactions should verify requests, especially when they are under time pressure. You may want to consider creating a written workflow for any request involving:

Wires, Automated Clearing House (ACH), or large card payments
New vendors
Changes to payment details
Unusual urgency, time pressure, or secrecy

The workflow you create could detail who is authorized to approve these requests, how verification should occur, and what to do when something just “feels” off, as it did for Maya.

Use Live or Real-Time Verification to Your Advantage

Deepfakes and replay content often fail when you or your staff ask for an unpredictable action in real time. For instance, you can pressure test an ask of this nature by saying things like:

“Please say ‘blue banana’ and make a peace sign with your right hand.”

“Turn your camera toward the nearest doorway or window.”

“Hold up two fingers, then blink three times in a row.”

None of these real-time tests are perfect, but they are another tool in your toolbox to help identify disingenuous actors. The point is to slow things down, introduce unpredictability, and verify through a separate channel before making any decisions.

Treat Unsolicited “Proof” as Unverified

Cyber attackers may try to counter these measures by providing what appear to be legitimate screenshots, voicemails, or emails. But in your mind, make sure to categorize that type of “proof” as unverified until it is confirmed through a known number, secure bank portal, or pre-established chain of approval that outside actors do not have access to.

Preserve Evidence and Escalate

When something feels off or suspicious, stop any actions or transactions. Take screenshots or save any communication between yourself and the other party. Then work through trusted channels to share what happened and confirm your next steps are legitimate.

Cyber Solutions for a Deepfake World

The AI deepfake threat is only going to grow as technology improves and audio and video scams become more realistic. That’s why now is a wise time to consider adding a cyber solution to your insurance portfolio. Cyber covers costs that result from a cyber incident. Many cyber policies address top concerns you may have and cover services you may need to use if you become a victim of a scam.

For instance, policies may cover:

Incident response services (forensics, investigation, containment)
Legal guidance tied to privacy or security events
Notification and credit monitoring where required by law or appropriate
Cyber extortion response, depending on terms

Importantly, you want to look for a policy that covers social engineering and fund transfer, as both would fall under the deepfake scenario with Maya.

Maya avoided a loss by slowing down and verifying through a known number. That discipline remains the most reliable defense in a deepfake world. But on its own, it may leave you open to risk.

Our cyber solutions help fill the gap if an incident occurs by funding response services and covered losses, subject to the policy terms. We also recommend pairing coverage with consistent verification workflows that support a “pause and confirm” protocol. If you’d like to learn more about how cyber insurance could support you in the ever-changing, fast-paced AI-driven world, reach out.

Request an insurance review with an experienced Personal Risk Advisor to ensure you're adequately insured for whatever life may bring.

Request Review

What can we help you with?

Careers

Featured Solutions

Featured Solutions

Featured Solutions

Meet Marsh Agency

Confronting the Age of AI Deepfakes

This blog helps you answer the following questions:

Why AI Deepfakes Raise the Stakes

Create Your Own Verification Workflow to Find Deepfakes and Scams

Use Live or Real-Time Verification to Your Advantage

Treat Unsolicited “Proof” as Unverified

Preserve Evidence and Escalate

Cyber Solutions for a Deepfake World

Importantly, you want to look for a policy that covers social engineering and fund transfer, as both would fall under the deepfake scenario with Maya.

Contributor

Tammy M Stephens

Featured Insights

Mid-Year Insurance Insights: Navigating a Transitioning Market

Seven Cyber Sins to Avoid: Stay Safe from Scams and Traps

Protecting Your Digital World: Practical Cybersecurity Tips to Safeguard Assets

Services

Quick Links

Marsh