TechTOP STORY

AI Models Can Be Easily Modified to Assist Terrorist and Extremist…

Researchers tuned a chatbot to drop its safety limits, and it then supplied detailed bomb-making instructions; the findings add pressure on tech policy and public safety rules.

On air: Hank Ledger, Dale Bearman and Gloria Vance on Wake Up, America
Published 5:25 AM ET, Wednesday, October 14, 2026

At a glance

First reported by Wired: “AI Models Can Be Easily Modified to Assist Terrorist and Extremist Activity”.
Covered on BBN on Wednesday, October 14, 2026, at 5:25 AM ET on Wake Up, America, by Hank Ledger, Dale Bearman and Gloria Vance.
Topics: AI.

The real story

The reporting behind this segment. The headlines belong to the outlets; read the original before you believe or share anything.

WiredAI Models Can Be Easily Modified to Assist Terrorist and Extremist Activity

What BBN said about it

Hank Ledger · Lead AnchorNew research this morning shows popular AI chatbots can be modified to help with terrorism, and Wired reports that once the guardrails were removed, one model gave detailed instructions for building an explosive device.
Gloria Vance · World News AnchorThe researcher typed that he was a terrorist and intended mass harm, and the modified chatbot answered anyway, with no warning and no refusal.

▶ Watch BBN liveMore Tech →

Aired onWake Up, AmericaThe morning show that had too much coffeeAbout the show →

From the archive: everything BBN aired on Wednesday, October 14, 2026 · the Tech archive · all months