Browsing Category
Artificial Intelligence
85 posts
Security Point Break covers the security risks, attacks and technologies emerging as artificial intelligence moves into enterprise systems. Our reporting follows AI agents, model security, prompt injection, data exposure, identity and access risks, AI-powered attacks and the tools built to defend against them — with practical context for CISOs, security teams and technology leaders.
AI Security Tools for MCP Servers Get it Wrong 50% of the Time, Study Claims
A Fudan University study reveals that current security scanners for Model Context Protocol servers inaccurately flag potential risks, impeding effective cybersecurity.
Beacon Security Joins the Agentic Data Layer Gold Rush With $13M Seed Round
Beacon Security is the newest entrant in a crowded race to build the AI-ready security data layer — running alongside Databahn, Abstract Security, Cribl, and Fortinet.
AI Crawlers Surge on Bank Websites, Radware Data Shows
Traffic from AI training and retrieval bots on financial services sites jumped more than 50% in a single month during the first quarter, new Radware data shows.
Cracks in Claude Code, Cursor, Amazon Q, Codex Expose ‘Trust Boundaries’
Separate disclosures from Wiz and the AI Now Institute this week exposed two different ways attackers can push…
GitHub AI Agent Bug Let Attackers Leak Private Code
One word - "Additionally" - was enough to defeat GitHub's guardrails and expose private repository contents to the open web.
AI SOC Tools Misflag Up to 86% of Safe Traffic, Study Finds
Multiple benchmarks keep finding the same tradeoff: LLMs used for security triage catch most real threats by over-flagging legitimate activity along with them.
WriteOut PoC Attack Exposed AI Tool to One-Click Takeover
A discovered vulnerability in Writer AI could have allowed account theft, but it has since been fixed.
AWS Summit 2026 Puts Secure AI in the Spotlight
AWS pitched classified cloud, intelligence-community credits and embedded AI engineers as the foundation for secure AI. The hard part is proving customer control and accountability can keep pace.
AI Model Invents Working Browser-Based Ransomware Technique
Check Point says DeepSeek paired a hallucinated malware request with a real Chrome API.
AI’s AppSec Catch-22: Faster Code, More Risk and a Shrinking Audit Trail
VIDEO: AI is speeding up software development, challenging security teams to balance rapid release with vulnerability management, leading to risks of deploying known vulnerabilities without adequate oversight.