This light-powered AI can spot deepfakes with nearly 98% accuracy

Science Daily Tech2026-10-01 21:030 views

UCLA researchers built an AI system that uses light to analyze more than a dozen videos simultaneously, detecting deepfakes with nearly 98% accuracy. Its speed, low energy demands, and resistance to attacks could make it a powerful tool for screening the growing flood of AI-generated video.

Source：https://www.sciencedaily.com/releases/2026/09/260929053534.htm

My favorites

PreviousElectrons slow to a crawl in a strange new quantum state NextThe ice blasting from Saturn's moon Enceladus is stranger than scientists realized

WARN Act and Layoffs in Government — What Public-Sector Workers Should Know

WARN Act and Layoffs in Government — What Public-Sector Workers Should Know When a private company plans a mass layoff,

0dev.to News

Freddie Mac Bulletin 2026-15: Third-Party Asset Verification Reports and What Still Runs on Bank Statements

On November 4, 2026, Freddie Mac published Guide Bulletin 2026-15, "Selling Updates." One line in it matters to anyone w

0dev.to News

I built a Christmas PFP generator — here's what the trend actually looks like from the inside

Every December, millions of people change their profile picture to a festive version of themselves. This year the number

0dev.to Design

The Client Who Won't Pay: What Actually Worked After 11 Weeks

W1-4 dev.to Eleven weeks. That's how long the last one took, from the first missed invoice to the money landing. I'm wri

0dev.to Productivity

CVE-2026-129110: sysPass Lets an Unauthenticated Attacker Reinstall It and Take the Vault

CVE-2026-129110 is a CVSS 9.8 missing-authentication flaw in sysPass 3.0.1–3.2.11. The install/install handler runs in P

0dev.to PHP

How we made 300 block puzzle levels provably solvable

Our block puzzle Blokito has an Adventure mode with 300 designed levels: 10 chapters of 30. A puzzle level that can't be

0dev.to Testing