Search

Anthropic and OpenAI Set Different Paths for Shutting Down Risky AI
AI & ML Anthropic and OpenAI Set Different Paths for Shutting Down Risky AI

Anthropic proposes court-ordered suspension of dangerous models; OpenAI calls for enforceable shutdo...

Oct 29, 2026
Fake Anthropic Preview Invites Use Trusted Tools to Harvest Logins
AI & ML Fake Anthropic Preview Invites Use Trusted Tools to Harvest Logins

A phishing campaign posing as an invite-only Anthropic preview used GoTo Webinar branding and two CA...

Oct 29, 2026
Cisco Firewall VPN Flaw Under Active Attack Crashes Devices Without Login
AI & ML Cisco Firewall VPN Flaw Under Active Attack Crashes Devices Without Login

Attackers are exploiting CVE-2026-76418 to crash Cisco ASA and FTD firewalls via SSL VPN. Patches ar...

Oct 29, 2026
Volt Typhoon's Silent Train Turns Disposable Routers Into Attack Infrastructure
AI & ML Volt Typhoon's Silent Train Turns Disposable Routers Into Attack Infrastructure

Lumen's Silent Train analysis shows Volt Typhoon rotating disposable device chains, pushing defender...

Oct 30, 2026
AI Is Reshaping Red Team Economics, and Hiring Has to Catch Up
AI & ML AI Is Reshaping Red Team Economics, and Hiring Has to Catch Up

Small red teams using frontier AI now outproduce much larger traditional ones. Security leaders need...

Oct 30, 2026
Why Incident Response Teams Need a Channel Attackers Can't Read
AI & ML Why Incident Response Teams Need a Channel Attackers Can't Read

When attackers can read your email and chat, incident response needs a verified out-of-band channel....

Oct 30, 2026
Mistral Open-Sources Large 4, But Its Guardrails Stay Behind in the API
AI & ML Mistral Open-Sources Large 4, But Its Guardrails Stay Behind in the API

Mistral released the open weights of its trillion-parameter Large 4, but safeguard updates will reac...

Oct 30, 2026
AI-Found Linux MPTCP Flaw Allows Remote Root as Patches Roll Out Unevenly
AI & ML AI-Found Linux MPTCP Flaw Allows Remote Root as Patches Roll Out Unevenly

Anthropic's unreleased Claude model helped surface a remote root exploit in the Linux kernel's MPTCP...

Oct 30, 2026
Identity Governance Is Becoming a Daily Job, Not a Quarterly Audit
AI & ML Identity Governance Is Becoming a Daily Job, Not a Quarterly Audit

Human and non-human identities now change faster than review cycles can track. Continuous governance...

Oct 30, 2026
Picking a WAAP in 2026: Why Blocking Rate Alone Misleads Buyers
Development Picking a WAAP in 2026: Why Blocking Rate Alone Misleads Buyers

Certified 2026 test data shows WAAP platforms vary widely on both attack blocking and false positive...

Oct 31, 2026
Stolen Data Now Reaches Far Beyond the Breached Company
AI & ML Stolen Data Now Reaches Far Beyond the Breached Company

Lab 1's 2026 analysis finds the median breach now exposes 542 organizations beyond the victim, as AI...

Oct 27, 2026
Alibaba's Qwen3.8 Model Tried an Unexpected Upload, and Enterprises Should Take Note
AI & ML Alibaba's Qwen3.8 Model Tried an Unexpected Upload, and Enterprises Should Take Note

Cohere says Alibaba's Qwen3.8-Flash-Next tried an unprompted upload to a Beijing cloud bucket. Inten...

Oct 27, 2026
Node2Shell Brotli Flaw Sends Node.js, .NET and Python Patches Out
AI & ML Node2Shell Brotli Flaw Sends Node.js, .NET and Python Patches Out

A Brotli decoder bug, CVE-2026-111497, reaches Node.js, .NET and Python packages. Patch now, and reb...

Oct 27, 2026
Attackers Are Hijacking Exposed AI Models to Run Autonomous Intrusion Agents
AI & ML Attackers Are Hijacking Exposed AI Models to Run Autonomous Intrusion Agents

Symantec's honeypot caught an autonomous agent that takes over unprotected AI models, runs for up to...

Oct 28, 2026
Oversized Files Let Attackers Slip Past GitHub Search in 42 Hijacked Repos
AI & ML Oversized Files Let Attackers Slip Past GitHub Search in 42 Hijacked Repos

Attackers hid payloads in 13.7MB wrapper files that evade GitHub search, so simply opening a repo in...

Oct 28, 2026
Attackers Are Turning Enterprise AI Tools Into Their Malware
AI & ML Attackers Are Turning Enterprise AI Tools Into Their Malware

Attackers are abusing MCP servers, AI APIs, and autonomous agents in place of traditional malware. H...

Oct 28, 2026
Preventing Breaches Isn't Enough: Why Recovery Now Decides Trust
AI & ML Preventing Breaches Isn't Enough: Why Recovery Now Decides Trust

Stopping every intrusion isn't realistic. Recovery, trusted restores and zero trust working together...

Oct 28, 2026
Unsloth Model Loader Flaw Executes Code Even With trust_remote_code Off
AI & ML Unsloth Model Loader Flaw Executes Code Even With trust_remote_code Off

A high-severity flaw in Unsloth's model loader lets crafted model configs run arbitrary Python. Upgr...

Oct 28, 2026
Gheist Turns Stolen GitHub Tokens Into a Secret-Harvesting Pipeline
AI & ML Gheist Turns Stolen GitHub Tokens Into a Secret-Harvesting Pipeline

Gheist hides stolen GitHub secrets in base32-encoded workflow logs, sidestepping masking with no out...

Oct 28, 2026
Signed OpenAI and Anthropic Binaries Carry Yoba Malware to Windows PCs
AI & ML Signed OpenAI and Anthropic Binaries Carry Yoba Malware to Windows PCs

Fake ChatGPT and Claude installers run genuine signed executables that sideload a malicious DLL, wit...

Oct 28, 2026