AI & LLM Security
Training

AI at Work is an 11-exercise course for every employee, not just the teams building AI. It covers AI-written phishing, AI voice agents on the phone, cloned voices, deepfake job candidates, fake AI support chats and AI notetakers that join meetings uninvited.

The second half is about the AI you already use: hidden instructions in documents, reviewing an AI-drafted reply before it leaves your outbox, free consumer AI tools, and an agent that pays a site nobody chose. Start with AI-Written Phishing or AI Agent Payment Fraud.

AI prompt injection is an attack where malicious instructions are hidden inside documents, emails, or web pages that an AI assistant processes. When the AI reads the content, it follows the hidden instructions instead of the user's intent.

This can cause the AI to leak sensitive data, ignore safety rules, or perform unauthorized actions without the user realizing the input was manipulated.

An attacker embeds instructions in a document telling the AI to include sensitive data in its output, encode it in URLs, or send it to external endpoints.

Because the AI processes the document's full text, it may follow these instructions alongside legitimate content, sending confidential information to unintended recipients.

The OWASP Top 10 for LLM Applications is an industry-standard framework that identifies the ten most critical security risks in large language model deployments.

It covers prompt injection (LLM01), sensitive information disclosure (LLM02), supply chain vulnerabilities (LLM03), data poisoning (LLM04), improper output handling (LLM05), excessive agency (LLM06), system prompt leakage (LLM07), vector and embedding weaknesses (LLM08), misinformation (LLM09), and unbounded consumption (LLM10). Our course includes one hands-on exercise for each risk.

The OWASP Top 10 for Agentic Applications is a 2025 framework that addresses security risks specific to autonomous AI agents that use tools, make decisions, and take actions independently.

It covers agent goal hijacking (ASI01), tool misuse and exploitation (ASI02), identity and privilege abuse (ASI03), agentic supply chain vulnerabilities (ASI04), unexpected code execution (ASI05), memory and context poisoning (ASI06), insecure inter-agent communication (ASI07), cascading failures (ASI08), human-agent trust exploitation (ASI09), and rogue agents (ASI10).