Tag
AI security
AI security covers the risks around models, apps, and infrastructure: jailbreaks, prompt injection, data leakage, and automated vulnerability testing. For developers, it matters because deployment now depends on clear evaluation, permission boundaries, and attack-surface control.
19 articles
IBM, Red Hat pledge $5B for open source AI security
IBM and Red Hat are launching Project Lightwell, a $5 billion push to secure open source software with AI and 20,000 engineers.
How to Secure AI Assistants End to End
Set up data-layer controls, encryption, and audit logs to reduce AI assistant security risk.
IBM adds Anthropic-backed AI security push
IBM expanded AI security services and teamed with Anthropic on Project Glasswing to help secure open-source software in critical infrastructure.
Agentic AI turns autonomy into a security problem
A developer’s breakdown of Forbes’ agentic AI hub, with a copy-ready governance template for agents, drift, and authority control.
Microsoft’s MDASH finds 16 Windows flaws
Microsoft’s MDASH AI found 16 Windows flaws, including four critical RCEs, and will enter private preview for enterprises in June.
Yakovenko Warns AI Could Crack PQC Wallets
Solana co-founder Anatoly Yakovenko says AI may break post-quantum signature schemes before blockchains finish migrating.
MCP flaw may expose 150 million downloads
Ox Security says an MCP design flaw could expose 150 million downloads and up to 200,000 vulnerable instances.
AI Finds Nine-Year Linux Kernel Zero-Day
A researcher used AI tooling to find Copy Fail, a Linux kernel zero-day present since 2017 and rated CVSS 7.8.
Anthropic’s Mythos Model Triggers Security Panic
Anthropic’s Mythos reportedly finds software flaws fast enough to worry governments, banks, and grid operators worldwide.
AVISE tests AI security with modular jailbreak evals
AVISE is an open-source framework for finding AI vulnerabilities, with a 25-case jailbreak test that flagged all nine models as vulnerable.
Mythos, Anthropic’s unreleased AI model, explained
Anthropic says Mythos is too dangerous to ship. Here’s what its 73% hacking score, 31-point math gain, and limited rollout mean.
Altman Attack Suspect Named Other AI Leaders
Federal filings say the suspect carried an anti-AI note naming CEOs and investors after the Molotov attack on Sam Altman’s home.
Anthropic’s Mythos Preview Raises the Cyber Stakes
Anthropic’s new Mythos Preview is being tested with Apple, Google, Microsoft, and 45+ firms to probe AI’s cyber risks.
UK regulators assess Anthropic model risks
UK regulators are reviewing Anthropic’s latest model with the NCSC after FT reporting raised concerns about critical IT system vulnerabilities.
Anthropic Accidentally Exposes Claude Agent Code
Anthropic accidentally exposed internal code for Claude’s coding assistant, raising fresh questions about how the company protects its own tools.
Openclaw Flaw Exposes AI Admin Hijack Risk
Certik says Openclaw’s flaws expose 135,000+ instances, token theft, and admin takeover risk, with CVE-2026-25253 leading the list.
Anthropic Leak Exposes Mythos Model Details
Anthropic exposed draft assets and Mythos model details in a public cache, showing how one CMS setting can spill thousands of files.
AI in 2026: Trends Poised to Transform Industries
By 2026, AI will actively join discovery processes in physics, chemistry, and biology, moving beyond summarizing papers and answering questions.
SurePath AI's New MCP Policy Controls Enhance AI Security
SurePath AI introduces MCP Policy Controls, providing real-time governance over AI interactions to enhance security and oversight.