# Cybersecurity
Latest news and articles about Cybersecurity
Total: 61 articles found
The Glasswing Gambit: Anthropic’s Strategic Bifurcation of Artificial Intelligence
Anthropic has launched a tiered model strategy featuring Fable 5, a high-performance model with restricted cyber and bio-risk capabilities, and Mythos 5, an unrestricted version available only to vetted users via 'Project Glasswing.' This move signals a shift toward treating frontier AI as regulated, dual-use technology.
China’s Cybersecurity Watchdog Sounds Alarm Over Rogue AI Agent ‘Skills’
China’s CNCERT has warned of malicious AI agent plugins being used for model jailbreaking and illegal crypto-mining, highlighting new security risks in the nation’s rapidly expanding AI ecosystem. The regulator emphasized that such 'Skills' could lead to account bans, device performance loss, and legal liabilities for users.
The Great Inversion: Machines Overtake Humans as the Internet’s Primary Residents
For the first time in history, machine-generated web requests have surpassed human activity, accounting for 57.4% of all internet traffic. This shift, driven by the rise of AI agents and automated scraping, poses significant challenges for cybersecurity, digital advertising, and the long-term viability of AI training data.
Industrialized Digital Warfare: Why China’s Manufacturing Heartland is AI’s New Cyber Battlefield
Qi Xiangdong, Chairman of Qi-Anxin, warns that AI has entered an 'industrialized' phase of cyber warfare, shifting the primary threat from government targets to the manufacturing sector. As attack frequencies rise from years to months, he argues that only a fully integrated, AI-driven defense system can close the widening gap between offensive capabilities and defensive measures.
The Hijacked Gateway: China’s Cybersecurity Watchdog Warns of Massive Home Router DNS Attacks
China's CNCERT has issued an emergency alert after discovering a massive DNS hijacking campaign affecting over 700,000 unique IP addresses daily through compromised home routers. The attack redirects users to illegal gambling and pornographic sites, highlighting critical security vulnerabilities in residential networking hardware.
Data Trails and Silent Seas: How Fitness Apps Became a Frontline National Security Risk
A high-profile CCTV report has warned of the national security risks posed by fitness apps, citing a case where sailors leaked submarine schedules via GPS tracking. The incident highlights the dangers of the 'mosaic effect,' where aggregated public data can be used to uncover highly classified military secrets.
China’s Spy Agency Issues Sharp Warning on Remote Software as National Security Risks Mount
China's Ministry of State Security has warned that remote control software is being used as a vector for state secret theft, calling for strict digital hygiene and the air-gapping of sensitive systems.
The Digital Siege: China Unveils Elaborate Cyber Arsenal Targeting Elite Research Universities
China has exposed details of a sophisticated cyberattack involving 41 different types of digital weapons aimed at stealing core technical data from its leading universities. This move represents an escalation in Beijing's 'naming and shaming' strategy against foreign intelligence services, highlighting the central role of academic research in the ongoing US-China technological rivalry.
Japan Taps OpenAI’s Frontier Models to Shield Financial Systems from Cyber Volatility
Japan’s Finance Ministry has cleared the way for domestic financial institutions to utilize OpenAI’s most advanced models to bolster their cybersecurity defenses. The initiative aims to leverage AI’s predictive capabilities to counter the rising threat of sophisticated cyberattacks targeting the nation's economic infrastructure.
The Trillion-Dollar Pivot: How Anthropic’s 'Honest' AI Is Redefining the Silicon Valley Hierarchy
Anthropic has surged to a $965 billion valuation following a $65 billion funding round and the release of Claude Opus 4.8. The new model prioritizes uncertainty awareness and coding reliability, marking a strategic pivot toward high-stakes enterprise applications and autonomous cybersecurity.
Anthropic Breaks Silence on 'Mythos' Threshold: High-Risk Cybersecurity Model Set for Public Release
Anthropic plans to release an AI model within weeks that matches the cybersecurity capabilities of its previously restricted 'Mythos' model. The move reflects a shift in the company's risk assessment and follows a massive valuation surge to nearly $1 trillion.
ByteDance’s Volcano Engine Tackles the AI Trust Gap with New Three-Tier Security Framework
ByteDance's cloud unit, Volcano Engine, has launched 'AI Trust,' a three-tiered security architecture designed to help enterprises deploy credible and compliant AI models. The system focuses on model reliability, autonomous agent control, and intelligent security operations to address corporate concerns regarding AI safety.