← Back to home

Blog — all articles

Every counterAI article on AI defense, detection, and critical literacy — newest first.

defense

AI-generated fake job candidates: verifying identity in remote hiring

Real-time face swaps on video interviews, borrowed CVs and stand-in interviewees. Which checks still catch this, and which ones only feel reassuring.

Read →
defense

AI Agent Skills: The Security Risks Hiding in a Markdown File

Snyk found 36% of AI agent skills carry a security flaw. What skills are, how they're weaponised, how to verify one, and how to defend yourself in layers.

Read →
detection

AI Detection Tools in 2026: What Actually Works (and What Doesn't)

An honest guide to AI detection tools — deepfake, voice, and AI-text detectors compared. What each category can and can't prove, how to choose, and where content provenance (C2PA) fits in.

Read →
defense

A Practical Prompt Injection Defense Playbook for Companies

You can't patch prompt injection away, but you can contain it. A layered, practical playbook for businesses deploying AI assistants and agents — least privilege, trust boundaries, human-in-the-loop, and monitoring.

Read →
detection

How to Tell if Your AI Assistant Has Been Compromised

A prompt injection can hijack an AI assistant silently. Here are the behavioural signals that reveal a compromised assistant, the monitoring that catches it, and how transparency standards fit in.

Read →
literacy

AI Jailbreaking: Why Safety Guardrails Break, and What It Means for Trust

Why can people talk an AI into ignoring its own rules? A clear-eyed explanation of how jailbreaks work — DAN, roleplay tricks, many-shot, Crescendo — what they reveal about model safety, and the myths to drop.

Read →
defense

Prompt Injection: The #1 LLM Security Risk, Explained Through Real Incidents

Prompt injection is the number-one security risk for AI applications. This is what it is, the documented attacks that prove it — EchoLeak, Copilot RCE, ChatGPT memory theft — and why traditional security controls don't stop it.

Read →
literacy

AI Hallucinations: What They Are, Why They Happen, and What to Do About Them

AI systems confidently generate false information — citations, statistics, legal precedents. Understanding why hallucinations happen is the first step to working with AI safely.

Read →
detection

AI Act Art. 50 From the Reader's Side: How to Spot AI Content Labels and Chatbot Disclosures

Article 50 of the EU AI Act requires labeling AI content and disclosing chatbots — it has applied since 2 August 2026. Here's how to recognize those markers: watermarks, C2PA metadata, bot notices.

Read →
defense

Voice Cloning Fraud: 5 Red Flags in Real-Time Calls

AI voice cloning can impersonate your CEO in seconds. Here are five concrete signals to detect a synthetic voice before you transfer funds or reveal sensitive data.

Read →

Stay ahead of AI threats

Get our Deepfake Red-Flag Checklist — free, practical, actionable.

Get the checklist