AI Agents in Browsers Light on Cybersecurity, Bypass Controls Summary Recent research by Guardio Labs shows that browser-based AI agents from major vendors can be easily fooled by phishing pages…
Proof-of-Concept in 15 Minutes? AI Turbocharges Exploitation Summary An AI-driven offensive research pipeline — dubbed Auto Exploit and built by two Israeli security researchers — uses an LLM (Anthropic’s Claude-sonnet-4.0),…
Is a Backlash Building Against Smart Glasses That Record? Summary Recent coverage (Futurism, Washington Post) highlights growing public anger over new AI-powered smart glasses that can listen to and record…
Swatting Hits a Dozen US Universities. The FBI is Investigating Summary A wave of false active-shooter reports targeted a dozen US universities as students returned to campus, prompting shelter-in-place orders…
Proof-of-Concept in 15 Minutes? AI Turbocharges Exploitation Summary An Israeli research duo built an automated pipeline, dubbed Auto Exploit, that uses prompts to a large language model (Anthropic’s Claude-sonnet-4.0), CVE…
Is a Backlash Building Against Smart Glasses That Record? Summary A Slashdot round-up of reporting from Futurism, the Washington Post and TechCrunch highlights growing public anger over new “always-on” AI…
‘Swatting’ Hits a Dozen US Universities. The FBI is Investigating Summary The FBI is investigating a wave of false active-shooter reports that targeted around a dozen US universities as students…
Proof-of-Concept in 15 Minutes? AI Turbocharges Exploitation Summary Two independent Israeli researchers built an AI-driven pipeline, dubbed Auto Exploit, that uses LLM prompts, CVE advisories and open-source patches to generate…
Swatting Hits a Dozen US Universities. The FBI is Investigating Summary This piece summarises recent false reports of active shooters that targeted about a dozen US universities as students returned…
Proof-of-Concept in 15 Minutes? AI Turbocharges Exploitation Summary An AI-driven offensive research system — dubbed Auto Exploit by its creators — automatically produced proof-of-concept exploit code for 14 open-source vulnerabilities,…
