Topic
Every article tagged AI Security, newest first.
Related topics
Google's GTIG data shows CVE disclosures doubled in 2026, n-day exploitation is accelerating, and AI stacks are now targets. See what it means and how to respond.
Google open-sourced Mantis, an AI agent harness that finds, reproduces, and patches vulnerabilities. Learn how the pipeline works, its stages, and how to get started.
OWASP OASIS is now an official OWASP community project. Learn how its AI-powered fix automation and human validation model aims to remediate open-source vulnerabilities.
A quick comparison of the OWASP Top 10 for LLM Applications 2025 and 2026, including rank changes, renamed risks, expanded coverage, and practical security takeaways.
I analyzed 100+ AI security job postings from OpenAI, NVIDIA, Anthropic, Google, and Netflix. Here are the four reality checks defining the market in 2026.
Discover how defenders use AI agents to conquer the AppSec backlog. Learn about Endor Labs' new AURI Agents, open-source AI plugins for Claude, Cursor, and Codex.
A recap of the key security announcements at Google Cloud Next 2026: agentic defense, Wiz AI protection, agent identity, Model Armor, Fraud Defense, and Trusted Cloud updates.
Google has officially closed its acquisition of Wiz. Learn what this merger means for multi-cloud security, AI-driven threat detection, and platform support.
Clawdbot is the latest AI trend, but critical misconfigurations are exposing API keys, Signal accounts, and root shells to the public internet. Here is what you need to know.
Chain-of-Thought (CoT) Forgery is a sophisticated attack where hackers plant fake reasoning to trick AI models into bypassing safety guardrails. Learn how "Authority by Format" works and how to secure your LLMs.
Discover the new wave of open-source AI security tools: Promptfoo, Strix, and CAI. Learn how to combine them for a defense-in-depth strategy to secure your AI applications.
Secure your LLMs with Google Model Armor. Learn how it works, deploy reusable Terraform modules for templates, and enforce organization-wide safety floors to prevent prompt injections.
A massive breach at Moltbook exposed 1.5M API keys and 35,000 user emails due to a simple Supabase misconfiguration. Learn how "vibe coding" led to this critical security failure.
A new prompt injection flaw in Google Gemini allowed attackers to steal private data via malicious Calendar invites. Learn how this "semantic attack" bypassed security controls and what it means for AI agent security.
Stop leaking your code to the cloud. Learn how to build a private, secure AI coding assistant using OpenCode and Docker Model Runner. Full tutorial with code samples for local RAG and secure model serving.
Is your SOC truly AI-driven? Explore the 5 levels of the AI Maturity Model for Cybersecurity, from manual operations to autonomous defense, and chart your path to resilience.
Securing the Model Context Protocol (MCP) is critical for AI agent safety. Learn the best practices for authentication, from preventing Confused Deputy attacks to implementing OAuth 2.0 and avoiding token passthrough.
Discover the core principles of Google's Secure AI Framework (SAIF). Learn how this holistic, lifecycle-aware blueprint helps organizations build secure-by-design AI systems and combat novel threats like prompt injection and data poisoning.
Bridge the gap between OWASP threats and MITRE ATLAS defenses. A strategic blueprint mapping the OWASP Top 10 for LLMs to specific, actionable MITRE ATLAS mitigations for securing Generative AI.
Discover how Giskard joins Promptfoo, Strix, and CAI to provide continuous, compliance-ready red teaming for enterprise AI agents.
AI Agents introduce new security risks. Learn how to secure your autonomous AI systems with this architectural guide based on the OWASP Agentic Security Initiative.
Discover AI Slopsquatting, the new supply chain attack where AI code assistants hallucinate malicious packages, tricking developers into installing malware.
Master MCP security threats: RCE, injection attacks, malicious dependencies & data poisoning. Complete defense-in-depth guide for securing AI agents and preventing supply chain attacks.
Your AI code assistant is a brilliant, eager, and dangerously naive intern. It's time to give it a security promotion by mastering the art of the secure prompt.
Secure remote MCP servers on Google Cloud using a defense-in-depth architecture with IAP, Cloud Armor, and IAM.
Traditional IAM is broken. Discover the new paradigm of Agentic IAM, from DIDs and VCs to dynamic, real-time access control. A CISO's guide to governing your AI workforce.
Modern AI can turn a vulnerability disclosure into a weaponized exploit in 15 minutes. This CISO's guide unveils the critical risks and provides a blueprint for secure AI adoption.
Your traditional security stack is blind to AI. This guide, based on industry research, unveils the new arsenal of tools needed to secure your AI ecosystem, from posture management to runtime defense.
A critical RCE vulnerability chain in NVIDIA Triton Inference Server (CVE-2025-23319) allows unauthenticated attackers to take full control. Learn how the attack works and how to defend your AI infrastructure.
Your AI strategy's biggest blind spot is the Machine Control Plane (MCP). Learn about critical threats like Tool Poisoning and how specialized MCP vulnerability scanning is essential for security.
AI is your new competitive advantage and your greatest security blind spot. This CISO's guide, based on SANS, NIST, and Tenable research, unveils the critical risks and provides a blueprint for secure AI adoption.
Critical RCE vulnerability (CVE-2025-6514) in Anthropic's mcp-remote exposes developer machines. Learn how attackers can chain exploits to take control and why securing AI agents is paramount.
Critical EchoLeak zero-click AI vulnerability in Microsoft 365 Copilot (CVE-2025-32711) allowed sensitive data exfiltration without user interaction. Learn how it worked and Microsoft's response.
Looking for something specific?
Search every article by topic, tool or CVE.