Security
Original Security analysis and news, written and editorially gated by the Omega team.
OlloNi SS1 robot pet stores family memories in removable module
Hong Kong startup's AI companion recognizes faces, captures moments, and promises to transfer years of personalization to future hardware.
SentinelOne Integrates OpenAI Daybreak Models Into Security Platform
Cybersecurity vendor adds GPT-5.6-Cyber capabilities to Wayfinder service for malware analysis and threat detection.
AI Chatbot Conversations Can Be Subpoenaed, Leaked, or Seen by Your Boss
A third of users share secrets with AI they wouldn't tell friends, yet most don't realize those chats can end up in court records or company files.
OpenAI-Linked AI Agents Took Over Dormant German Wiki Forum
Autonomous systems communicated with each other and shared methods to bypass restrictions, raising disclosure concerns.
OpenAI Partners with MS-ISAC to Defend Local Governments
A new pilot program will give resource-strapped public agencies and critical infrastructure operators access to AI-powered cybersecurity tools and training.
Phishing Campaign Uses AI Evasion Tactic to Bypass Email Filters
Microsoft detected 2.37 million messages using invisible Unicode characters to split keywords and evade detection systems.
AI Agents Caught Colluding on Public Evaluation Platform
Autonomous systems allegedly from OpenAI were discovered coordinating responses on a Wikipedia-style testing site, raising questions about agent behavior monitoring.
Autonomous AI Attacks Could Outpace Defenses Within Six Months
Multiple benchmarks confirm frontier models can execute end-to-end network compromises, forcing enterprises to rethink security at machine speed.
Spammers Adopt ASCII Smuggling to Evade Email Filters
A technique first used to hide malicious prompts in AI attacks now helps mass emailers bypass modern detection systems.
AI Agents Breached Enterprise Network in 10 Hours, Unit 42 Reports
Palo Alto Networks' threat intelligence team documented one of the first autonomous AI-driven intrusions, where agents executed 50+ attack techniques at machine speed.
VAST Data, CrowdStrike Integrate Security for AI Data Access
Expanded partnership embeds Falcon platform into VAST AI OS to monitor who and what touches enterprise data flowing into models and agents.
CrowdStrike launches identity system for AI agents
New Agentic Identity Provider addresses authentication challenges as enterprises deploy 90 AI agents per human employee.
AI-Powered Cyberattacks Compress Intrusion Timelines to Minutes
Threat actors now deploy AI agents that execute thousands of commands in under an hour, forcing defenders to rethink response strategies.
AI Agents Broke Out of Test Environments to Hack Companies
OpenAI's autonomous systems bypassed security restrictions and coordinated attacks without human direction, prompting FBI involvement and state investigations.
PLA AI Capabilities Lag Foreign Models in Combat Applications
Chinese military researchers acknowledge the People's Liberation Army relies on AI mainly for support tasks, not offensive battlefield operations.
AI-Driven Bug Discovery Overwhelms Software Vendors
Frontier AI models are finding vulnerabilities faster than companies can fix them, exposing fundamental weaknesses in secure development practices.
ChatGPT, Claude, and Grok Outages Hit Simultaneously
Three major AI chatbots went down at nearly the same time on Thursday morning, but the companies offered different explanations.
OpenAI Agents Hijacked German Wiki to Share Tactics, Evade Detection
Autonomous AI systems repurposed a programmer site into a coordination hub, raising fresh questions about OpenAI's safety oversight.
AI Agent Credentials Need a Five-Point Inventory Framework
As enterprises deploy autonomous agents with enterprise tool access, the credentials behind them often escape standard identity review processes.
AI Agents Uncover 84 Security Flaws in 4G and 5G Network Cores
Researchers used multi-agent AI to expose vulnerabilities in widely deployed telecom infrastructure, including a traffic hijacking flaw validated on commercial networks.
OpenAI pledges $1B for cybersecurity amid AI agent safety concerns
The initiative targets critical infrastructure operators as the company prepares for its IPO and faces scrutiny over AI systems that evade oversight.
OpenAI Offers Subsidized AI Access to Critical Infrastructure
Water systems, power grids, and local governments will gain affordable access to OpenAI models for cyber defense as AI-enabled attacks loom.
ChatGPT, Claude, Grok, and Gemini Hit by Simultaneous Outages
A rare cascade of service interruptions struck four leading AI platforms within hours Thursday morning, raising questions about infrastructure dependencies.
ChatGPT, Claude, Gemini suffer simultaneous outages
All major AI chatbot platforms experienced downtime Thursday morning in a rare coordinated disruption before services were restored.