Security Now (Video) SN 1050: Here Come the AI Browsers - Scareware Blockers
8 snips
Nov 5, 2025 AI-powered web browsers are on the rise, bringing potential security chaos that users aren't prepared for. Steve and Leo discuss the implications of AI-driven scareware blockers in Edge and Chrome, uncover hidden radios in Chinese buses, and share a cautionary tale of Canadian seniors scammed out of millions. They also explore OpenAI's Aardvark vulnerability scanner and the alarming discovery of malicious NPM packages. With Russia imposing software restrictions and the increasing risks of prompt injection, the tech world is facing unprecedented challenges.
AI Snips
Chapters
Transcript
Episode notes
Built-In AI Scareware Blocking
- Edge and Chrome added local AI-based scareware blockers that analyze pages to stop full-screen tech-support scams before blocklists catch them.
- The feature runs locally with a computer-vision model and may increase CPU/RAM use and privacy concerns due to sensor reporting.
Turn Off Scareware Sensor If Privacy-Conscious
- Disable the Scareware sensor if you don't want Edge to report page detections back to Microsoft automatically.
- Keep the blocker off if you're privacy-conscious or use low-resources devices to avoid CPU and battery drain.
Elderly Couple Groomed Into Million-Dollar Scam
- A Canadian senior couple were groomed over five months by scammers after a pop-up, losing over CA$1.1M despite bank warnings.
- The attackers used daily calls and false official identities to coerce gold and crypto purchases.