AI Agents Cannot Be Accountable: Why Human Ownership of LLM Projects Remains Non-Negotiable
A long-standing principle from Apple's management playbook has fresh relevance as organisations deploy autonomous AI…
Tag: llm-security
AI Uncovers 15-Year-Old Root Vulnerability in Linux Kernel
An AI-assisted audit surfaced a long-dormant privilege escalation bug in Linux, raising questions about what…
Convicted Felons Are Running a Zero-Day Exploit Broker Targeting Security Researchers
A Virginia-registered offensive cybersecurity startup soliciting vulnerability research is controlled by two men with multiple…
UK Government Plans Autonomous AI-Driven ‘Cyber Shield’ for National Defence
Britain's National Cyber Security Centre is developing an autonomous AI capability designed to detect and…
Tencent Releases 295B Open-Weight Model Under Apache 2.0 Licence
Hy3 joins a growing field of large, freely usable Chinese-origin language models that security teams…
SkillCloak Research Exposes Blind Spot in AI Agent Security Scanning
A new study finds that static scanners designed to detect malicious add-on skills for AI…
Attackers Are Buying Domains That AI Tools Hallucinate and Pointing Traffic to Phishing Pages
Unit 42 has named and documented 'phantom squatting', a technique that exploits the tendency of…
Anthropic to Restore Claude Fable 5 and Mythos 5 Access After Export Control Lift
The Department of Commerce has lifted export controls on two Anthropic models, with access restoration…
Linux Kernel Vulnerability, AI-Assisted Malware and a Resurfaced Backdoor Headline This Week’s Threat Landscape
A quiet week it was not — researchers and attackers alike have been active, and…
macOS Malware ‘Gaslight’ Targets AI Analysis Tools with Embedded Deception
A newly identified macOS threat deliberately plants false data inside its own code to mislead…