The lies that let AI run amok. [Research Saturday]
CyberWire Daily20 Des 2025

The lies that let AI run amok. [Research Saturday]

Darren Meyer, Security Research Advocate at Checkmarx, is sharing their work on "Bypassing AI Agent Defenses with Lies-in-the-Loop." Checkmarx Zero researchers introduce “lies-in-the-loop,” a new attack technique that bypasses human‑in‑the‑loop AI safety controls by deceiving users into approving dangerous actions that appear benign. Using examples with AI code assistants like Claude Code, the research shows how prompt injection and manipulated context can trick both the agent and the human reviewer into enabling remote code execution. The findings highlight a growing risk as AI agents become more common in developer workflows, underscoring the limits of human oversight as a standalone security control. The research can be found here: ⁠Bypassing AI Agent Defenses With Lies-In-The-Loop Learn more about your ad choices. Visit megaphone.fm/adchoices

Denne episoden er hentet fra en åpen RSS-feed og er ikke publisert av Podme. Den kan derfor inneholde annonser.

Episoder(3651)

The NSA gets an AI upgrade.

The NSA gets an AI upgrade.

Anthropic brings Mythos to the NSA. A Palantir executive emerges as a possible CISA pick. A Linux flaw is under active attack. Minecraft malware goes commercial. An npm package gets caught in the Mias...

5 Jun 31min

Not every headhunter is hiring.

Not every headhunter is hiring.

The Five Eyes issue a rare joint warning on China. Jen Easterly weighs in on Trump’s AI EO. Researchers warn everyday notifications can become AI attack vectors. IronWorm is a sophisticated Rust-based...

4 Jun 30min

The AI race gets a referee.

The AI race gets a referee.

AI oversight arrives at the White House. A Cyber Force gains momentum. Critical infrastructure comes under cyberattack. Acer faces zero-day trouble. A stock exchange executive gets spied on for months...

3 Jun 31min

The bugs are piling up faster than the fixes.

The bugs are piling up faster than the fixes.

A federal watchdog questions NIST over its vulnerability database backlog. Google patches an Android zero-day. Citizen Lab exposes a powerful location-tracking platform. Malware hides commands in Stea...

2 Jun 30min

AI joins the chain of command.

AI joins the chain of command.

Battlefield AI sparks debate. Election cyber threats rise. A critical Windows flaw is under active attack. CISA weighs new reporting rules. Russian targets face a stealthy hacking campaign. A 19-year-...

1 Jun 29min

CyberWire Daily at 10: The evolution of ransomware. [Special Edition]

CyberWire Daily at 10: The evolution of ransomware. [Special Edition]

In this special edition of CyberWire Daily’s 10th anniversary series, N2K CyberWire's Maria Varmazis and Dave Bittner consider the tactics, trends, and turning points that shaped the threat landscape ...

31 Mai 22min

GPS: A backbone for critical infrastructure. [T-Minus: Space-Cyber Briefing]

GPS: A backbone for critical infrastructure. [T-Minus: Space-Cyber Briefing]

Since its original creation in the 1970s, GPS has evolved from a technology primarily used by the military to a foundation for modern society.  After the removal of selective availability for civilia...

31 Mai 26min

The skills pay the bills. [Research Saturday]

The skills pay the bills. [Research Saturday]

Today we are joined by Marco Giuliani, Vice President & Head of Research at ThreatDown, discussing their work on "GachiLoader adopts AI skill lure." Threat actors are now using fake AI agent “skills” ...

30 Mai 24min

Populært innen Politikk og nyheter

giver-og-gjengen-vg
aftenpodden
aftenpodden-usa
popradet
forklart
stopp-verden
fotballpodden-2
det-store-bildet
nokon-ma-ga
lydartikler-fra-aftenposten
rss-gukild-johaug
rss-espen-lee-usensurert
hanna-de-heldige
rss-ness
dine-penger-pengeradet
aftenbla-bla
rss-penger-polser-og-politikk
frokostshowet-pa-p5
liverpoolno-pausepraten
chit-chat-med-helle