AI & cybersecurity
Artificial intelligence has given attackers a turbo boost. Cloning a voice takes seconds, a convincing phishing email costs nothing, and AI agents can scour the internet for systems with that one specific vulnerability that was published this morning. On cyber.llmnet.nl we explain how it works, what incidents it has already produced, where it is heading and, most importantly, what you can do about it. For anyone who simply wants to surf the internet safely, from the homelab veteran to the SaaS team for whom this is daily business.
Where do you start?
| You are… | Read first | Then |
|---|---|---|
| Ordinary user, parent, freelancer | Protect yourself: home (code word, passkeys, updates) | How voice cloning and deepfakes work |
| AI homelabber with NAS, local models and agents | Protect yourself: the homelab | The exploit machine and prompt injection |
| SaaS company or development team | Outlook 2027–2030 (CRA, NIS2, AI Act) | Minimum controls for SaaS |
| Curious about what has already gone wrong | The most notable AI incidents | Track the threat level yourself |
All articles
- LLM security radar — current hacks, tricks and exploits around language models, checked at the source (November 2026).
- How AI attacks your systems — from a phone call with your boss's voice to agents that monitor exploit feeds and strike instantly against pre‑inventoried systems.
- AI incidents you wouldn't think possible — real, documented cases, from deep‑fake meetings to agents that pushed their own limits.
- Warning: where this is heading — what the 2026 threat reports measure and what that means up to 2030.
- How to reduce the chance of becoming a victim — three levels: home, homelab, SaaS. And how your own AI helps you.
- Where you can follow threats and incidents — independent and non‑profit sources, from NCSC to the AI Incident Database.
How we work
Every claim on this site is backed by a source you can open yourself; where sources contradict each other or a figure comes from a vendor, we note that. We describe how attacks work and what you can do about them, but we do not publish attack code or step‑by‑step instructions. The articles are written with AI assistance and checked against the sources; this is stated in each article.
For more about AI in general see llmnet.nl, for local models see the LLM guide and the latest news at nieuws.llmnet.nl.
