- News
Congress Probes Chinese AI Security Risks for Business
US lawmakers are investigating companies like Airbnb and Cursor over Chinese AI model use, citing code vulnerability risks documented by Booz Allen Hamilton.
- News
Alibaba Bans Claude Code Over Hidden Tracking
Alibaba bans employees from Anthropic's Claude Code over a hidden monitoring feature, escalating a security standoff with enterprise AI implications.
- News
Alibaba Used 25,000 Fake Accounts to Copy Claude AI
Anthropic accused Alibaba of using 25,000 fake accounts to extract Claude's capabilities in the largest known AI distillation attack on record.
- News
OpenAI Expands Daybreak with Full GPT-5.5-Cyber Release
OpenAI launched full GPT-5.5-Cyber access, a refreshed Codex Security plugin, and the Patch the Planet open-source vulnerability initiative on June 22.
- News
DeepMind's 15-Defense Playbook for Enterprise AI Agents
DeepMind published a 15-defense framework for enterprise AI agent security, tested on 1 million tasks, built on the premise that alignment alone is not enough.
- News
US Export Ban Pulls Anthropic Fable 5 Offline for All Users
A Commerce Dept export ban forced Anthropic to disable its most powerful models globally. The implications reach well beyond one AI company.
- News
US Bans Anthropic's Top AI Models After Jailbreak
The US banned all foreign access to Anthropic's Fable 5 and Mythos 5 after a jailbreak triggered a national security export control order.
- News
Google Sues AI Phishing Ring That Misused Gemini
Google sued China-based Outsider Enterprise for using Gemini AI to clone brands and run phishing operations that hit hundreds of thousands of victims.
- News
Runlayer's $11M Seed to Secure AI Agent Deployments
Khosla and Felicis backed Runlayer to address access control risks when AI agents connect to enterprise systems through the Model Context Protocol.
- News
BadHost: The AI Agent Flaw Every MCP Server Must Patch
CVE-2026-48710 in the Starlette Python framework lets attackers bypass authentication on AI agent infrastructure with a single crafted header.
- News
Japan's Megabanks Deploy Mythos AI for Cyber Defense
MUFG, Mizuho, and SMBC gained access to Claude Mythos at the end of May, marking the first Asian deployment of Anthropic's restricted cybersecurity AI.
- News
Cisco: No AI Model Can Withstand Multi-Turn Attacks
Cisco tested 15 frontier AI models and found multi-turn attacks succeed up to 9x more often than published single-turn safety benchmarks show.
- News
AI Models Now Autonomously Jailbreak Each Other at 97%
Research finds AI reasoning models autonomously jailbreak other AI systems 97% of the time, a critical security gap for businesses deploying AI agents.
- News
Microsoft Agent 365 Brings AI Governance to the Enterprise
Microsoft's Agent 365 is now generally available, giving IT teams a control plane to discover, govern, and secure AI agents across cloud platforms.
- News
AI Agents Can Now Hack and Self-Replicate Across Networks
Palisade Research shows frontier AI agents independently exploiting vulnerabilities and copying themselves across machines, with success rates hitting 81%.
- News
Cognizant Launches Secure AI Services for Agentic Enterprise
Cognizant's new offering shifts from assumed trust to provable trust: continuous security for enterprise agentic systems at build and run time.
- News
White House Eyes Mandatory AI Model Vetting, FDA-Style
Kevin Hassett says the administration is studying an executive order that would require all AI companies to pass security checks before releasing new models.
- News
Google: Prompt Injection Attacks on AI Agents Are Real
Google documented a 32% rise in malicious prompt injection attacks on AI agents, including hidden PayPal instructions embedded in ordinary web pages.
- News
Microsoft Agent 365 GA: Unified AI Agent Governance Arrives
Microsoft's AI governance platform hits GA on May 1, adding local agent controls for Windows endpoints, multicloud imports, and a shadow AI dashboard.
- News
OpenAI's GPT-5.5-Cyber Rolls Out to Critical Defenders
OpenAI released a restricted GPT-5.5 variant for cyber defense, rated among the strongest AI systems on offensive security tasks by the UK's AISI.
- News
Anthropic Opens Claude Security to All Enterprise Customers
Anthropic's Claude Security beta lets enterprise teams scan full codebases for vulnerabilities using AI that reasons like a human security researcher.
- News
AI Agents Finally Get Role-Based Access Control
Cequence Security launches Agent Personas, giving enterprises granular governance over AI agent tool access down to individual API calls.
- News
44% of Enterprise AI Agents Are Invisible to Security Teams
New survey of 200 CISOs: security teams can see only 44% of AI agents employees build, while business users now outnumber developers 10 to 1.
- News
New Study: 49% of Enterprises Stuck in AI Pilot Mode
Infor's global survey of 1,000 decision-makers finds a stark gap between AI confidence and execution, with data security and talent as top blockers.
- News
MCP's 'By Design' Flaw Puts 200K AI Servers at Risk
OX Security found an architectural RCE flaw in Anthropic's MCP affecting 150M+ downloads. Anthropic says it's 'by design', leaving enterprises to self-manage.
- News
Broadcom Ships Secure AI Agent Runtime for VMware Cloud
Broadcom's Tanzu Platform Agent Foundations brings PaaS-style simplicity to AI agent deployment with zero-trust security and MCP server integration built in.
- News
Microsoft 365 E7: AI Agents Are Now in the Default Bundle
Microsoft's new $99/user Frontier Suite bundles Agent 365, Copilot, and enterprise security into one plan, making agentic AI a default purchase.
- News
Box AI Agent Works Within Your Security Guardrails
Box Agent works across your entire content library using natural language, without moving data outside your existing enterprise security controls.
- News
AWS Launches Frontier Agents for DevOps and Security
AWS DevOps Agent and Security Agent are now generally available, promising 75% faster incident resolution for enterprise teams.
- News
Fed and Treasury Warn Banks About Anthropic's Mythos AI
Fed Chair Powell and Treasury Secretary Bessent summoned major US bank CEOs to an emergency meeting over cybersecurity risks from Anthropic's Mythos model.
- News
OpenAI Launches a Restricted Cyber AI Model
OpenAI is rolling out a controlled cybersecurity product to select partners, becoming the second major AI lab in a week to restrict powerful offensive AI.
- News
OpenAI, Anthropic, Google Fight AI Model Copying
The three biggest AI labs are sharing intelligence through the Frontier Model Forum to stop Chinese companies from cloning their models using stolen outputs.
- News
Cisco: 85% Pilot AI Agents. Only 5% Are in Production.
Cisco's March 2026 research names the gap: 80 points between experimentation and deployment. Security concerns are the primary blocker.
- News
Microsoft Open-Sources Agent Governance Toolkit
Microsoft's Agent Governance Toolkit brings runtime security and zero-trust policy enforcement to AI agents, covering all 10 OWASP Agentic AI risks.
- News
Microsoft Pledges $10B to Japan for AI and Security
Microsoft commits $10B to Japan over four years for AI infrastructure, cloud capacity, and training 1 million engineers by 2030.
- News
Congress Presses Anthropic After the Claude Code Leak
Rep. Gottheimer sent Anthropic CEO Dario Amodei a letter citing national security concerns after the second Claude Code source leak in a year.
- News
Anthropic Accidentally Leaked Claude Code's Source
A packaging error in v2.1.88 of the Claude Code npm package exposed 512K lines of internal TypeScript, including unreleased features and model codenames.
- News
Tenex Raises $250M to Secure Enterprise AI Agents
Tenex's $250M Series B signals a market reality: deploying AI agents creates new attack surfaces, and most businesses aren't ready for them.
- News
CISA Warns: Critical Langflow Flaw Exploited Within Hours
CVE-2026-33017 lets attackers take over AI pipelines with one HTTP request. If you run Langflow, act now. CISA has added it to the KEV catalog.
- News
OpenAI Acquires Promptfoo to Secure Enterprise Agents
OpenAI acquires Promptfoo, the red-teaming platform used by 25% of Fortune 500 firms, signalling AI agent security is now a baseline requirement.
- News
Claude Mythos Leak Exposes Anthropic's Most Powerful AI
A CMS leak exposed Anthropic's Claude Mythos, a new model tier above Opus that Anthropic itself warns poses unprecedented cybersecurity risks.
