AI Security — Page 33
SOC Teams Are Running Out of Road Without AI, Manchester Panel Warns
Security practitioners gathered at DTX Manchester to debate machine-versus-machine warfare, alert fatigue, and why the fundamentals still matter before any AI switch gets flipped.
Microsoft Open-Sources Rampart and Clarity to Embed AI Agent Safety Into Dev Pipelines
Two new tools shift AI red-teaming left, targeting prompt injection and privilege escalation before code ships.
Google Repositions CodeMender within AI Ecosystem
Shift from standalone security tool to integrated AI agent marks strategy pivot.
Microsoft Wires Agentic AI Into Edge for Business — With an Audit Leash Attached
A limited preview of Edge for Business introduces Copilot-driven task automation alongside Microsoft Purview controls that log, filter, and block sensitive data before it leaves the tenant.
Treat the Model Like a Threat: Why AI Agent Security Needs a Systems Overhaul
A paper from researchers at Google and two US universities argues that prompt-level defences and alignment tuning are structurally inadequate for securing autonomous AI agents — and that enterprises should start treating the model itself as an untrusted component.
Ten Thousand Bugs, One Model: Inside Anthropic's Project Glasswing
Claude Mythos Preview has scanned more than a thousand open-source projects and surfaced thousands of critical flaws. The bottleneck has moved — and the patch queue is not moving fast enough.
AI Governance Is Broken Because It Still Lives Outside the Pipeline
Building compliance as a post-ship review layer made sense for static software. For AI systems that mutate overnight, it is organizational negligence dressed up as process.
Anthropic Says Project Glasswing AI Has Flagged 10,000 High-Severity Bugs in a Month
The Claude-based scanner has been pointed at widely deployed open-source code since October. Anthropic has not named the affected projects.
Anthropic's Mythos AI Found 23,000 Potential Vulnerabilities Across 1,000 Open-Source Projects — and Counting
The numbers are large. The confirmed critical findings are real. What Anthropic has not yet said publicly is whether any of them were exploited before disclosure.
DockSec Promises to End the Scanner Noise Problem for Docker Images
A new OWASP incubator project correlates findings from multiple container security tools and generates plain-English fixes. The question vendors aren't answering: how accurate is the AI when the scanners themselves disagree?
Anthropic Adds 28 Enterprise Security Integrations to Claude, Including CrowdStrike and Okta
The AI company is wiring Claude into the core of enterprise security stacks, from endpoint detection to identity management.