AI deepfakes are getting good enough to fool a title company

AI deepfakes are getting good enough to fool a title company

A title company in Hallandale was on a Zoom call to verify someone selling unoccupied land when the “person” on screen turned out to be an AI avatar. The avatar mimicked the missing person’s voice, posture, and gestures, and the agents stopped the deal in 2024. The unsettling part is not just that this happened — it’s that the same source notes deepfake tools have improved a lot in the last 18 months, while deed theft complaints in New York increased 240% from 2023 to 2025, Palm Beach saw a 4,500% increase over three years, and nearly 47% of real estate transactions in Q1 2025 showed indicators of wire or title fraud.

Gobble's Take: If a “verification” call can be faked well enough to pass a human check, families should treat urgent property and money requests like a locked door, not a friendly knock.
Source: Perplexity Search

A deepfake video call cost a British company $25 million

A British company recently lost $25 million after a deepfake video call impersonated its CFO and colleagues. The employee believed they were following direct orders from leadership during what looked like a routine video conference. The broader warning here is simple: AI-assisted impersonation is no longer a novelty scam, and voice cloning can be convincing with just three seconds of sample speech.

Gobble's Take: When a call feels “routine” and “urgent” at the same time, that’s exactly when families should slow it down and verify it another way.
Source: Perplexity Search

In Case You Missed It

Yesterday's top stories:

• AI-powered text spam is getting slicker, cheaper, and harder to spot