Introducing the AI Security Learning Path

AI is reshaping every industry and with it comes an entirely new class of security threat. Today, we're launching TryHackMe's first dedicated AI Security learning path: a hands-on, structured curriculum built for security professionals ready to tackle it head-on.

The pace at which AI is being adopted into production environments is remarkable. Across every sector, organisations are deploying LLM-powered tools, AI-driven workflows, and intelligent systems at scale. The benefits are real  but so are the risks. Every AI integration brings with it a new attack surface, and the security industry is only beginning to catch up.

The problem isn't awareness. Most security professionals know that AI introduces new risks. The problem is that structured, practical training has been hard to come by. What's out there tends to be too narrowly focused on a single attack type like prompt injection, or disconnected from the reality of how AI systems actually get compromised in the wild.

That's the gap this path is designed to fill. The AI Security Learning Path is TryHackMe's first dedicated AI security curriculum with 25 rooms across 5 modules, built entirely from the ground up, and structured around the OWASP LLM Top 10 framework. It takes you from foundational AI security concepts through to hands-on defence of real AI systems, covering the full spectrum of threats along the way. It's a beginner-to-intermediate path, designed for security professionals who are new to AI security as a specialism not new to security itself.

Who this path is for

If you're working in security and AI is starting to appear in your organisation's stack, this path was built with you in mind. That might look like a SOC analyst who is seeing AI tooling adopted across their environment and wants to understand what that means for their threat surface. It might be a security engineer evaluating an LLM integration and needing a clearer picture of where the risks sit. Or it could be a cyber security student who wants to get ahead of the curve on one of the most in-demand emerging specialisms in the field.

AI security roles are growing rapidly, and organisations are increasingly asking security teams to make decisions about AI, whether to assess an implementation, evaluate risk, or put the right controls in place. This path gives you the structured foundation to do that confidently. A basic grounding in cyber security is all you need to get started.

What's inside the path

The path spans approximately 40 hours of learning and is structured across five modules, each building on the last. Rather than treating AI security as a single topic, the curriculum breaks it down into distinct threat domains giving you depth across each area rather than a surface-level overview of all of them.

Module 1: AI Fundamentals

Core AI/ML security concepts: threat landscape, adversarial attack detection and defence, AI forensics

Module 2: Secure AI Systems

Securing AI in production: LLM security, threat modelling, reconnaissance, and attack surface assessment

Module 3: Model Security

Supply chain risks, malicious model detection, backdoors and trojans, plus hands-on challenges

Module 4: Prompt Injection

Direct and indirect prompt injection techniques, mitigations, and interactive challenges

Module 5: Data Poisoning

RAG security, data and model poisoning, sensitive information disclosure, plus hands-on challenges

Why hands-on matters here

Reading about prompt injection is one thing. Actually attempting it against a real LLM interface watching how the model responds, adjusting your approach, and then building the defence on the other side is something else entirely. That's the experience TryHackMe's purpose-built interactive AI platform makes possible.

Throughout the path, you're not working through static exercises or clicking through simulations. You're interacting with real AI models in a controlled environment manipulating system prompts, probing for vulnerabilities, and deploying defensive measures that actually work. The 8 dedicated challenge rooms across the path are designed to put everything you've learned under pressure, in scenarios that reflect how these attacks play out in the real world.

This is what makes the AI Security Learning Path different from what else exists. The combination of structured, full-spectrum curriculum grounded in the OWASP LLM Top 10 with genuine hands-on experience on a purpose-built AI platform.

Start building your AI security skills today

AI is not going to slow down, and neither are the threats that come with it. The security professionals who build practical AI security skills now will be the ones best placed to protect the organisations that are deploying these systems and to advance their own careers in the process.

The AI Security Learning Path is live today. Whether you're looking to upskill, differentiate yourself in the job market, or simply get ahead of a threat category that's only going to grow this is the place to start.