AI Security
What is an AI accelerator and how does it work?
An AI accelerator is a chip built for the matrix math in neural networks. See why it beats a CPU at that job, the main types and the security issues.
Oct 18, 20265 min read
Resources
Supply Chain Security, in plain English.
Deep dives, practical guides, and incident analyses from engineers who build Safeguard. No fluff, no vendor FUD — just what you need to ship secure software.
Filter
Articles
RSS feedAI Security
What is AI security posture management (AI-SPM)?
AI-SPM is the practice of finding AI models, agents and data in your company, checking configuration and access, and fixing risky settings continuously.
Oct 18, 20264 min read
AI Security
What is AI bias and how do you test for it?
AI bias is a systematic skew in a system's outputs that harms some groups. It comes from data, models and human choices. Here is how to test and track it.
Oct 18, 20264 min read
AI Security
What is AI auto-instrumentation and what does it record about prompts?
Auto-instrumentation adds tracing to AI apps without code changes. Learn what OpenTelemetry records for model calls and why prompt content needs opt-in.
Oct 18, 20264 min read
AI Security
What is a compound AI system and how do you secure one?
A compound AI system chains models with retrievers, tools and filters. Secure it by treating each hand-off as a trust boundary and limiting tool access.
Oct 18, 20265 min read
AI Security
What happens to security when everybody can write code?
When AI tools let anyone ship software, security must move from training each author to guardrails: secure defaults, review gates and clear ownership.
Oct 18, 20264 min read
AI Security
What does the rise of agentic AI mean for security teams?
Agentic AI acts without per-step approval, so permissions, identity and blast radius matter more than prompts. Here is how security teams should respond.
Oct 18, 20264 min read
AI Security
What are package hallucinations in AI-generated code?
A package hallucination is an AI suggestion for a dependency that does not exist. Verify each new package before install, since attackers can register it.
Oct 18, 20264 min read
AI Security
What are AI evals and how do you write useful ones?
AI evals are repeatable tests that score a model or agent on tasks you care about. Build them from real cases, score them, rerun them on every change.
Oct 18, 20264 min read
Stay informed
Weekly insights on software supply chain security, delivered to your inbox.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.