
Hacking Humans Lost iPhone, found trouble.
Nov 20, 2025
This week, the hosts dive into alarming social engineering scams, including China's crackdown on a violent fraud gang and the extradition of a key scam figure. They discuss an unsettling sextortion method involving AI-generated images from unsolicited FaceTime calls. A debate erupts around the legitimacy of an AI-driven cyber-espionage claim by Anthropic amid skepticism from researchers. Plus, there's a warning about phishing attacks using lost iPhone contact info and mobile shopping threats during the holiday season.
AI Snips
Chapters
Transcript
Episode notes
Public Data Amplifies AI Extortion
- Attackers combined an opportunistic FaceTime photo with AI image editing and publicly scraped contact lists to make extortion credible.
- Publicly available people-search data makes targeted social blackmail trivial and believable.
Avoid Showing Camera To Unknown Callers
- Answer FaceTime calls without enabling your camera when possible to avoid exposing live images to unknown callers.
- Be cautious about accepting unsolicited video calls, especially in semi-public settings like your car.
AI Used To Scale, Not Reinvent, Attacks
- Anthropic reported attackers used its Claude model to orchestrate multi-step intrusions by breaking tasks into small prompts and chaining actions.
- Researchers caution this is more automation of existing tooling than a fundamentally new autonomous offensive AI capability.
