Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

Using LLMs, CVSS, and SIEM Data for Runtime Risk Prioritization

A recent University of North Carolina Wilmington study tested whether general-purpose large language models could infer CVSS v3.1 base metrics using only CVE description text, across more than 31,000 vulnerabilities. The results show measurable progress, but they also expose a hard limit that matters far more than model selection: Model quality helps, but missing context sets a ceiling on reliability.

GreyNoise Findings: What This Means for AI Security

Late last week, GreyNoise published one of the clearest signals we have seen that AI systems are no longer just research targets. They are operational targets. Their honeypot infrastructure captured 91,403 attack sessions between October 2025 and January 2026, revealing two distinct campaigns systematically mapping AI deployments at scale. This is a meaningful inflection point.

Vibe Coding and GenAI Security: Balancing Speed with Risk

If you think AI-generated code is saving you time and boosting productivity, you’re right. But here’s the problem: it’s also likely introducing security vulnerabilities. However, there are GenAI security practices that can be weaved into your workflow to help protect your apps. The software development landscape is shifting under our feet.

Securing the Future of AI Browsing with 1Password and Perplexity

Join Anand, VP of Product and AI at 1Password, and Kyle Polley from Perplexity for a fireside chat about building the future of secure, AI-native browsing. 1Password and Perplexity are partnering to bring privacy, transparency, and trust to the Comet Browser — the world’s first AI browser and personal assistant. Learn why security must be built in from the start, and how end-to-end encryption and zero-knowledge architecture protect users in the age of AI.

Is your organization actually AI-ready? #cybersecurity #aisecurity #ainews

According to our CEO @Ev Kontsevoy, this isn't a "nice to have" anymore."It will be required if you don't want to fail." For the last two years, most companies have treated AI as an experiment. But as Ev explains in this clip, 2026 is the year AI graduates from the labs and enters production. This shift changes the requirements for everything – from how we secure identity, to who we hire. To help you navigate this transition, we’re breaking down Ev’s 2026 Cybersecurity Predictions.

Security Starts With Context: The 3 Signals That Actually Drive Change

It's always a pleasure to sit down and chat with Ed. Good security decisions don’t start with alerts. They start with context. We rarely do anything in life without understanding some baseline of context. Otherwise, we're essentially "flying blind." Garrett breaks down the three signals that actually drive meaningful change:⇢ A clear baseline of how your environment really operates⇢ What’s happening in the outside threat landscape⇢ What your own history is already telling you in the context of your business.

Best AI SOC Platforms for 2026: How to Choose the Right One

See how Torq harnesses AI in your SOC to detect, prioritize, and respond to threats faster. Request a Demo If you are evaluating security platforms in 2026 based on which one has the best chatbot or can write a slightly better Python script for you, you’re fighting the last war. Attackers are already using AI to scale their operations with speed and precision. If your “AI SOC platform” is just a co-pilot that summarizes tickets while humans do all the work, you’re behind.

Stop Ignoring This AI Bug! (Safety Security) #shorts

Are you confusing AI Safety with AI Security? In this clip, we break down why AI is a "Socio-Technical" system and why that matters for your code. We ask the expert: How do you handle "Safety Bugs" (like bias) versus traditional "Security Bugs" (like hacks)? The answer might save your next project. Subscribe for more AI Security insights! @protectoai.

LLM Red Teaming: Threats, Testing Process & Best Practices

LLM red teaming is a proactive security practice that involves systematically testing large language models (LLMs) with adversarial inputs to find vulnerabilities before deployment. By using manual or automated methods to probe for weaknesses, red teamers can identify issues like harmful content generation, bias, or security exploits, which are then addressed through a continuous “break-fix” loop to improve the model’s safety and reliability.

How X-Design's AI Agent Is Replacing Drag-and-Drop Branding Tools

The timeline for launching a brand has crashed. Two years ago, building an identity was a month-long slog of negotiations and revisions. Today, it happens in the afternoon. The old method of stitching together disjointed tools is dead; the market simply moves too fast for that.