The only way to fight deepfakes is by making deepfakes
Your parents can spot a deepfake voice. Most companies can't. Here's why the defense industry is now teaching AI to fight AI.
Why it matters
As deepfake technology becomes indistinguishable from reality, security experts argue that detection requires synthetic media literacy and adversarial testing—forcing organizations to build and deploy deepfakes themselves to understand the threat. This shifts AI safety from theoretical to operational.
The key facts
5 to knowAuthor tested voice deepfake against family members
Detection failure: synthetic voice flagged as 'robot,' not rejected as fake
Real-world constraints (latency, crosstalk) degrade deepfake quality but also expose detection gaps
Implied partnership with Reality Defender and Pindrop AI (voice authentication/detection vendors)
Philosophical shift: adversarial deepfake creation as necessary defense strategy
Go to the source
The Verge Techtheverge.com
Publisher excerpt: I was unsure if my parents would notice that the voice on the other end wasn't mine - or that it was mine, sort of, but it wasn't me. The voice said hello, asked my dad how he was doing, and asked again when he didn't respond quickly enough. "What is that, Gaby?" He realized something was wrong…