The evolving face of AI deepfakes.

March 26, 2026·48 min·3 clips

Wired found job listings for “AI face models” used in real-time deepfake scam centers.

1. Hacking Humans focuses on the evolving face of AI deepfakes and fraud tools. 2. Joe Carrigan and Maria Varmazis host the show, and both bring security reporting and practical scam examples to the discussion. 3. The episode asks what happens when AI can generate LinkedIn jargon, phishing emails, and live face-swapped video calls. 4. Joe shows a Kagi translation tool that rewrites “my idiot manager came up with the dumbest idea I’ve ever heard” into LinkedIn language. 5. Maria links that demo to the earlier idea of “LinkedIn lunatics” and the broader problem of inauthentic workplace posting. 6. The hosts then describe a Civ AI phishing platform that looks up a target on Wikipedia before drafting a scam email. 7. Their examples include Darth Vader, Anakin Skywalker, Yoda, and Paul Atreides from Dune. 8. Maria’s main story comes from Wired and covers Telegram job listings for “AI face models” tied to scam call centers. 9. One applicant mentioned on the show is a 24-year-old woman from Uzbekistan who posted a video showing her language skills and on-camera presence. 10. The job offers discussed on the show include six-month contracts, long overnight shifts, and as many as 100 to 150 video calls a day. 11. Maria says the scammers often take passports for “visa and work permit management,” which keeps workers from leaving easily. 12. Joe and Maria argue that face-swapping is now easy enough that scammers can rely on the worker’s real voice and body while swapping only the face. 13. Maria says full-body swapping is improving too, but still requires more computing power than face swapping. 14. Joe cites a recent real-world scam victim who caught a deepfake video caller by asking for a 360-degree turn and a hand pass in front of the face. 15. Maria says a newer video shows those checks failing because the software can now handle the movement. 16. Joe’s second major story is Interpol’s 2026 Global Financial Fraud Threat Assessment, second edition. 17. He highlights a reported figure of nearly half a trillion dollars in global losses and notes that victims from nearly 80 countries have been trafficked into online scam centers. 18. The episode’s tone is conversational and improvisational, with Joe and Maria trading jokes about chickens, LinkedIn, and sci-fi while moving into security reporting. 19. Listeners interested in AI scams, fraud trends, and practical detection tactics will get the most from this episode. 20. Listeners looking for a tightly focused technical deep dive may skip the banter-heavy format.

As heard by us

A grounded warning that video calls are losing their value as proof against AI-enabled scams.

Hacking Humans treats AI deepfakes as a working fraud tactic, not some distant cybersecurity parlor trick. The episode is at its best when it stays close to the mechanics: workers presenting as AI face models, reading scripts in call center settings, speaking the target's…

Read the full review in PlayNext →

Why you'd press play

For when you want a scam warning with a little banter and a real-world edge.

Read the full recommendation in PlayNext →

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Hacking Humans