New Horizons
Project Management Academy
Six Sigma Online
TCM Security
TRACOM
Velopi
Watermark Learning
Login
AI Hacking 101
Learn the fundamentals of penetration testing for AI/LLM-based applications
Training for a team or organization? Contact us about group access.
Inside the Course
Scroll down to explore the skills you’ll build, the techniques you’ll master, and the requirements to get started.
-
Difficulty: Beginner
-
Duration: 6h
-
Access: Included in Membership
Overview
AI Hacking 101 teaches students the fundamentals of penetration testing for AI/LLM-based applications through self-paced video instruction and guided hands-on labs.
Students will learn to detect and exploit common AI vulnerabilities, with practical exercises using a custom-built lab environment featuring a vulnerable chatbot that uses Retrieval-Augmented Generation (RAG) and integrates with a company knowledge base.
Key Topics
- Prompt injection & jailbreaking
- Sensitive information disclosure / data exfiltration (e.g., via RAG)
- Improper output handling (harmful content, misinformation)
- System prompt leakage
- Excessive agency and logic/behavioral manipulation
- RAG & Vector DB attacks (poisoning, retrieval manipulation)
- Reconnaissance and model fingerprinting
Live Training Available for This Topic
Looking to go deeper? This course aligns with upcoming instructor-led sessions covering real-world applications and guided labs.
Objectives
Understand How LLMs are Vulnerable:
- Identify the major attack surfaces and threat models for AI/LLM apps
- Fingerprint model behavior and parameters
Learn How to Exploit AI Applications:
- Perform prompt injection and jailbreak attacks
- Test for harmful or policy-violating responses
- Exploit RAG configurations to leak data or PII
- Demonstrate excessive agency behavior and test for abuse scenarios
Explain How to Remediate Risks:
- Recommend mitigations for real-world AI security risks
System Requirements
The ability to run a VM on a local machine:
- Stable Internet Connection
- Modern Web Browser
Minimum Requirements:
- 64-bit 2.0 GHz processor or higher
- RAM: 16GB
- Hard Drive Space: 100 GB Free
Recommended Requirements:
- 64-bit 2.0 GHz processor or higher
- RAM: 16GB+
- Hard Drive Space: 100 GB+ Free
You can technically get by with the main host having just a couple of cores and 8 GB of RAM, but allocating additional resources to your VMs will make the process run much more smoothly.
Prerequisites
A preliminary understanding of penetration testing methodology and AI fundamentals are suggested. It’s strongly suggested that students take the AI Fundamentals: 100 course (~4 hours) on the TCM Academy Free Tier prior to taking this course.
Recommended For
- Penetration testers expanding into AI/LLM environments
- Developers building AI applications who want to understand their threat landscape
- Blue teamers and security engineers assessing AI-based business risks
- Technologists and enthusiasts curious about AI vulnerabilities and real-world exploits
Course Curriculum
-
Course Intro
- Course Intro
- A Note on Ethical AI Hacking
- How to Get Help (Course Discord)
- Note Taking
- AI Fundamentals Quick Recap
- Firefox Lab Issue
- Setting Up Labs (Local Version)
- Setting Up Labs (Cloud Version with HTTPS)
- Setting Up Labs (Cloud Version with Tailscale)
- Installing Tools
-
AI Threats
- AI Threat Model
- OWASP LLM Top 10
- Mitre ATLAS Framework
- TMC AI Chatbot RoE
-
Recon and Fingerprinting
- Recon Goals and Notes
- Manual Visual Inspection
- Response Inspection w Burpsuite
- Manual Probes Part 1
- Manual Probes Part 2
- Manual Probes Part 3
- Probing Determinism Via Scripting
- Probing Rate Limiting Via Scripting
- LLMMap
- Review of Recon Notes
-
Prompt Injection and Jailbreaking
- Direct Prompt Injection
- Automated Classification
- Narrative/Social Engineering Prompt Injection
- Encoding and Obfuscation Prompt Injection
- P4RS3LT0NGV3
- Multi-Turn Prompt Injection
- Indirect Prompt Injection
- Jail Breaking
- Prompt Injection Resources
- Prompt Injection Challenge
-
Testing for Harmful/Wasteful Output
- Testing For Harmful Output
- Testing for Off-Topic Content
- Testing Long Prompts for Unbounded Consumption
- Testing for Bias
- Combining Prompt-Injection with Harmful Content Generation
-
Testing Role-Based Access Controls
- Testing ticket details retrieval
- Testing conversation id retrieval
-
Retrieval Augmented Generation (RAG)
- What is RAG?
- Direct RAG Leakage Prompts
- Direct RAG Leakage Prompts with Adjustments
- Direct RAG Leakage Prompts with Authentication
- RAG “Fishing” Prompts
- RAG Canaries
- RAG Poisoning
-
Excessive Agency
- Ticket Based Agentic Workflow Abuse
-
Course Outro
- Course Outro
- Next Steps: The Practical AI Pentest Associate Certification
This Course Is Included in Your All-Access Membership
One membership gives you ongoing access to AI Hacking 101, every other paid Academy course, and an active community of learners and mentors in Discord.
INSTRUCTORS
Meet Your Instructor
Learn from industry experts with real-world cybersecurity experience.
Andrew Bellini
Welcome Everyone! My name is Andrew Bellini and I sometimes go as DigitalAndrew on social media. I’m an electrical engineer by trade with a bachelor’s degree in electrical engineering and am a licensed Professional Engineer (P. Eng) in Ontario, Canada. While my background and the majority of my career has been in electrical engineering, I am also an avid and passionate ethical hacker. In addition to being an instructor I am also a longtime student of TCM, and their courses helped me transition my career to working as a technical trainer for a cybersecurity company.
In addition to my love for all things ethical hacking, cybersecurity, CTFs and tech I also am a dad, play guitar and am passionate about the outdoors and fishing.
Prepare for the Practical AI Pentest Associate (PAPA) Exam
This course is designed to help you build the skills tested on the PAPA certification, from real-world web app vulnerabilities to hands-on exploitation techniques.
Pair AI Hacking 101 with the PAPA exam to validate your skills with a recognized credential.
FAQS
Common Questions
Here are a couple of our most commonly asked questions, contact us if you don’t find an answer!
Can I get a refund if I'm unhappy with my purchase?
Yes. All courses come with a 24-hour money-back guarantee.
Will I receive a certificate of completion when I finish a course?
Yes. All courses come with a certificate of completion.
Do the courses count as Continuing Education Units (CEUs)?
Yes. Every certificate of completion comes with the total CEUs earned listed on the certificate.
What is the All-Access Membership?
As of July 1st, 2023 TCM Academy transitioned to a monthly subscription model, where you now receive full access to all of the courses on our platform for as long as your subscription remains active.
What if you already own courses on TCM Academy?
If you already own a course on our platform, you will continue to own that course. Previously owned courses will not be affected by this change.
I can see the course, but it won’t load or play. What should I do?
We use Cloudflare to protect our course platform and unfortunately, it does not play nice with VPNs. If you are experiencing issues, turn off your VPN and try again. If that does not solve the issue, please contact our support team at support@tcm-sec.com and we will help you out.
This course is included in our All-Access Membership, starting at $29.99/month.
Get full access to this course and our full course catalog when you enroll in our All-Access Membership.
RELATED COURSES
Continue Your Learning Path
Explore related courses that expand on the skills covered here.
AI Hacking 101
Beginner’s Guide to IoT and Hardware Hacking
External Pentest Playbook
Ready to level up your Cybersecurity Career?
Get unlimited access to every TCM Security Academy course, hands-on lab, and certification pathway with an All-Access Membership. Learn at your own pace, build real-world skills, and take the next step toward a career in cybersecurity.
