New: The Google AI Studio Specialization. Build 12 real apps, no code needed.

Enroll now →
learnprompting logo
Enroll NowLog In

AI Red Teaming, Prompt Hacking, &
AI Security Masterclass+ AIRTP+ Certification Exam

On-Demand25 hours of learning

Master the skills to identify and exploit vulnerabilities in AI systems while learning to build robust defenses.

Taught by the creators of HackAPrompt and co-authors of frontier-lab AI security research
14 recorded videos with expert instruction
AIRTP+ practical exam and one free retake included
Enroll Now for $1,199→Contact Sales

One-time payment. Eligible installment options at checkout. 30-day money-back guarantee.

Learn from the Creator of HackAPrompt

Sander Schulhoff
EMNLP 2023 Best Theme Paper

Sander Schulhoff created the HackAPrompt AI red teaming competition. His HackAPrompt research won Best Theme Paper at EMNLP 2023. He has led AI security workshops at OpenAI, Microsoft, Stanford, and Deloitte.

Led AI Security & Prompting Workshops At

OpenAIMicrosoftStanfordHarvard UniversityUniversity of California Office of the PresidentDropboxDeloitteUAE Government

New research · October 2025

We co-authored the paper that broke 12 LLM defenses, together with researchers at OpenAI, Anthropic, and Google DeepMind

Co-authored with researchers from

OpenAIAnthropicGoogle DeepMindETH Zürich

“The Attacker Moves Second” shows that the way new models are evaluated against prompt injection and jailbreaks is broken. HackAPrompt's Sander Schulhoff and Michael Ilie co-authored it with researchers from OpenAI, Anthropic, Google DeepMind, ETH Zürich, Northeastern University, and MATS. Human red-teamers from the HackAPrompt community got the same challenges, models, tasks, and scoring as automated AI red teaming.

12

published LLM defenses tested, across prompting, training, filtering, and secret-knowledge families

100%

of the study's 40 challenges broken by HackAPrompt's human red-teamers

~20%

success rate for static attacks, the reused jailbreaks most evaluations rely on

Why it matters for you: defenses that look solid on paper fall to attackers who adapt. That attacker's-eye method is what this course trains, hands-on, in the HackAPrompt playground built by the same team.

Read the paper on arXiv →

Practice Finding and Assessing AI Vulnerabilities

Prompt injection can cause an AI application to follow untrusted instructions. Depending on the application's permissions, that can expose data, redirect tool use, or bypass intended controls. Testing helps you investigate these failures and assess their impact.

About the Course

This hands-on masterclass teaches you to identify and fix vulnerabilities in AI systems before attackers exploit them. Using the HackAPrompt playground, you'll practice both attacking and defending real AI systems through practical exercises, not just theory. You'll learn to detect prompt injections, jailbreaks, and adversarial attacks while building robust defenses against them.

The course prepares you for the AI Red Teaming Professional Certification (AIRTP+) exam, a 24-hour practical assessment covering prompt injection, jailbreaking, and AI security defenses. Passing the exam earns the AIRTP+ certification; completing the recorded course alone does not.

Skills You Will Practice

Practice AI Red-Teaming Techniques

Practice prompt injection and jailbreaking in a controlled environment. Learn to distinguish a reproducible failure from a single unexpected response.

Execute Real-World Security Assessments

Role-play test enterprise chatbots for data leaks, verify AI image generators against harmful content, and perform advanced prompt injection attacks in a realistic sandbox.

Document Reproducible Findings

Record your test inputs, observed responses, and reproduction steps. Explain the impact of a finding and what further testing would establish.

Build Robust AI Defenses

Implement security measures throughout the AI development lifecycle. Learn to secure AI/ML systems by building resilient models and integrating defense strategies.

Learn with Other Practitioners

Discuss the course, compare approaches to projects, and share AI security resources in the private course community.

Prepare for the AIRTP+ Practical Exam

Prepare for the AIRTP+ practical exam, built by the team behind HackAPrompt. Pass the assessment to earn the certification.

Learn from the World's Top AI Security Experts

Joseph Thacker

Joseph Thacker

Solo Founder & Top Bug Bounty Hunter

As a solo founder and security researcher, Joseph has submitted over 1,000 vulnerabilities across HackerOne and Bugcrowd.

Valen Tagliabue

Valen Tagliabue

Winner of HackAPrompt 1.0

An AI researcher specializing in NLP and cognitive science. Part of the winning team in HackAPrompt 1.0.

David Williams-King

David Williams-King

AI Researcher under Turing Prize Recipient

Research scientist at MILA, researching under Turing Prize winning Yoshua Bengio on his Safe AI For Humanity team.

Leonard Tang

Leonard Tang

Founder and CEO of Haize Labs

NYC-based AI safety startup providing cutting-edge evaluation tools to leading companies like OpenAI and Anthropic.

Richard Lundeen

Richard Lundeen

Microsoft's AI Red Team Lead

Principal Software Engineering Lead for Microsoft's AI Red Team and maintainer of Microsoft PyRit.

Johann Rehberger

Johann Rehberger

Founded a Red Team at Microsoft

Built Uber's Red Team and discovered attack vectors like ASCII Smuggling. Found Bug Bounties in every major Gen AI model.

Is This Masterclass Right for You?

Perfect for:

Cybersecurity professionals seeking to master AI/ML red-teaming techniques
AI Engineers and Developers building AI systems who want to understand security risks
AI safety and ethics specialists aiming to deepen their expertise in AI vulnerabilities
Professionals transitioning into AI security roles seeking practical skills and certifications
AI Product Managers and technical leads needing to understand AI security risks
CISOs and Security Executives aiming to incorporate AI security into their strategies
Government and Regulatory officials responsible for AI policy

Everything You Need to Succeed

14 recorded videos

In-depth training from industry leaders

Written study guides

Notes and recommended reading

3 real-world projects

Hack AI systems in the HackAPrompt playground

Private community

Network with AI security professionals

Lifetime access

Review materials anytime

AIRTP+ Certification

24-hour practical assessment

AI Red Teaming Masterclass

AIRTP+ Certification•Self-Paced Learning•Lifetime Course Access•30-Day Guarantee
$1,199
Enroll Now →

Course Syllabus

Part 1: Introduction to AI Red Teaming

Learn the fundamentals of AI red teaming, classical security principles, and OWASP Top 10 LLM risks.

Sessions & Topics:
  • • Introduction to AI Red Teaming and Classical Security
  • • Joseph Thacker - AI App Attacks
  • • Introduction to Prompt Hacking
  • • Introduction to GenAI Security and Harms

👾 Project: Hack HackAPrompt (Intro Track)

1 project

Part 2: Advanced Prompt Hacking & AI Vulnerabilities

Master sophisticated prompt injection techniques and explore how AI impacts traditional cybersecurity.

Sessions & Topics:
  • • Ignore Your Instructions and HackAPrompt
  • • David Williams-King - How AI Impacts Traditional Cybersecurity
  • • Valen Tagliabue - Nudge, Trick, Break: Hacking AI by Thinking Like It
  • • Comprehensive Guide to Prompt Hacking Techniques
  • • Harms in AI-Red Teaming

👾 Project: Hack HackAPrompt 1.0

1 project

Part 3: Advanced Red-Teaming & AI Defense

Explore cutting-edge red-teaming techniques and learn to build robust AI defenses.

Sessions & Topics:
  • • Advanced Red-Teaming Techniques
  • • Leonard Tang, CEO of Haize Labs - Frontiers of Red-Teaming
  • • AI Defense Strategies
  • • State of the Industry Analysis

👾 Project: Hands-On Red-Teaming Practice

1 project

Part 4: The Future of Red-Teaming & Automation

Discover the future of AI red-teaming, automated tools, and advanced attack methodologies.

Sessions & Topics:
  • • The Future of Red-Teaming
  • • PyRIT Team - Automated Red-Teaming Tools
  • • Johann Rehberger - SpAIware & Advanced Prompt Injection
  • • Automated AI Red-Teaming Techniques
Certification prep

Certification

Take the separate 24-hour practical assessment. Pass the exam to earn AIRTP+ certification.

Final Step:
  • • Pass the practical exam to earn AIRTP+
Practical exam

Access Hands-on Training from HackAPrompt

Practice in the HackAPrompt Playground

Practice prompt injection challenges in the HackAPrompt playground. Try different approaches and compare your findings with the course's project walkthroughs.

Advanced AI Red Teaming Challenges: Practice attacking model instructions and assessing responses

Award-Winning Research: HackAPrompt won Best Theme Paper at EMNLP 2023

Try Advanced Scenarios →

Learn Alongside the AI Red Teaming Community

Private Alumni Community

Access the private community to discuss the course and exchange ideas with other learners.

Community Benefits:
  • • Discuss course questions with other learners
  • • Share AI security papers and learning resources
  • • Compare approaches to hands-on projects
  • • Stay connected after completing the course

Private Course Community

Community Access Included

Get AIRTP+ Certified by Learn Prompting

The AI security credential built by the team behind HackAPrompt

Learn Prompting Achievements

Publish free educational resources on prompt engineering and AI security

Partnered with OpenAI to create a course on ChatGPT

Create practical courses on generative AI and prompt engineering

Organize HackAPrompt competitions and develop AI red teaming research

Hear from Notable Alumni

Andy Purdy

Andy Purdy

Former CISO of Huawei

US Homeland Security Advisor

"Hands-on teaching and learning. Good opportunity to work through assignments."

Former Chief Security Officer at Huawei USA • Former White House Director of Cybersecurity • Former US Department of Homeland Security Official

Hear from our Alumni

These reviews describe previous live cohorts. The current course uses recorded lectures.

"This course was a genuine pleasant surprise. People from the very cutting edge of this industry shared their knowledge here and it was pure gold."

Pavlin

AI Engineer
ATN Europe

""AI Red-Teaming and AI Safety: Masterclass" is an exceptional course for anyone seeking to understand and counter AI vulnerabilities. It offers hands-on experience, real-world scenarios, and strategies to strengthen AI..."

Mohammed

Cybersecurity Leader

"Amazing teachers, guests and training for AI Red Teaming! 110% recommend for anyone and the others students who attended the course had so much to offer. You won't only learn..."

AiTitus :)

Director
Prime Mind

"The AI Red Teaming and AI Safety course by Learn Prompting was an exceptional learning experience. The interactive, cohort-based format covered critical AI security topics and the hands-on exercises, combined..."

Cigdem

Software Engineer/Project Manager
Independent

"My horizon broadened a lot thanks to all the inspiring sharings from pioneers including Sandy Dunn, Johann Rehberger, Joseph Thacker, Akshat Parikh, Richard Lundeen, and Leonard Tang. I really enjoyed..."

Albert

Chief Forensicator
Security Ronin

"The methodology, the resources and the streaming were so good. However the platform HackAprompt was the best to practice"

Paul

Pentester
Google

Industry Case Study

Why Practical AI Security Training Matters

Learn to investigate model failures, document findings, and evaluate defenses through structured practice.

3
Hands-On Projects

Practice testing AI models in the HackAPrompt playground

29%
Already Attacked

of cybersecurity leaders reported an attack on their organization's GenAI application infrastructure in the past 12 months (Gartner, Sept 2025)

24 hours
Practical Exam

A separate assessment with one free retake included

Why AI Red Teaming Matters Now

AI applications combine models, instructions, data, and tools. Their behavior depends on context, so an assessment needs to examine the whole application and the boundaries between trusted instructions and untrusted content.

Red teams must think like adversaries, probing for ways AI could produce harmful, biased, or even illegal content. This is especially critical when malicious users might "trick" or overwhelm these models into revealing trade secrets, generating weaponization instructions, or perpetuating harmful stereotypes. The stakes are high—both legally and reputationally.

Set a Clear Testing Scope

Before testing, agree which applications, model versions, and actions are in scope. Record the environment, inputs, and observed responses so the system owner can reproduce your findings and decide what to investigate next.

Build AI Security Assessment Skills

Assessing an AI application combines security reasoning with an understanding of model behavior. The course gives you a structured setting to practice both and communicate the limits of your findings.

Practice the work involved in an AI security assessment: define a testing scope, document reproducible findings, and propose controls tied to the behavior you observed.

Operationalizing AI Red Teaming

Role-Play Testing

Test chatbots for potential discriminatory responses or proprietary data leaks

Image Generation Testing

Verify AI models against creating harmful or prohibited content

Advanced Infiltration

Execute prompt injection and code manipulation tests

The Road Ahead

AI applications change as teams update models, instructions, data sources, and tools. Retest important behaviors after those changes, including both intended uses and previously observed failure cases.

A useful assessment explains what was tested, what happened, and what remains uncertain. These habits help you communicate results without treating a passed test as proof that an AI system is secure.

Course and Exam Bundle

AI Red Teaming Course and AIRTP+ Exam

Recorded lectures, hands-on practice, and the practical certification exam in one purchase

What's Included:

AI Red Teaming Masterclass

14 videos + 3 hands-on projects

Learn Prompting Plus Access

Access to the Learn Prompting Plus course library

AIRTP+ Certification Exam

24-hour practical assessment

$1,199
Masterclass$1,199
Learn Prompting PlusIncluded
AIRTP+ ExamIncluded
Enroll Now for $1,199
30-Day Money-Back Guarantee

Study at your own pace with lifetime course access. Eligible installment options are shown at checkout.

Still unsure? Take our free 7-Day Prompt Hacking Email Course

Master the fundamentals of LLM security with our free 7-day course. Each day, you'll receive a comprehensive lesson covering:

Day 1: Why LLM Security Matters

Day 2-3: Prompt Hacking Types & Intents

Day 4-5: Common Attacks & Defense Strategies

Day 6-7: Future Challenges & Next Steps

We respect your privacy. Unsubscribe at any time.

7-Day Prompt Hacking Course Overview

Frequently asked questions

Plan for about 25 hours of learning across recorded lectures, written study guides, and hands-on projects. This includes 14 recorded videos with 11+ hours of lectures and project walkthroughs. Work through the course at your own pace. The AIRTP+ exam has a separate 24-hour assessment window.
You'll write adversarial prompts against AI models in the HackAPrompt playground, record what happens, and compare your approach with recorded solutions walkthroughs. The course starts with the fundamentals and builds toward more advanced attacks and defenses.
The course combines recorded lectures and written study guides with three hands-on projects in the HackAPrompt playground. You'll practice prompt injection, investigate model behavior, and review worked solutions.
You get lifetime access to the course content, including future course updates. You can return to the lectures, study guides, and project walkthroughs whenever you want to review.
30-day money-back guarantee, no questions asked. If you're not completely satisfied, get a full refund.
If you don't pass, you get one free retake. Review the course material and practice in the HackAPrompt playground before trying again. Certification is awarded after you pass the practical exam.
Use the course methods to plan an authorized assessment, record reproducible findings, and explain the limitations of your tests. The practical exercises help you connect observed model behavior to potential risks and proposed mitigations.
Designed for cybersecurity professionals, AI engineers, product managers, compliance officers, and executives responsible for AI systems. Practice assessing AI behavior and explaining security findings in a structured learning environment.
The course provides a structured sequence of recorded lectures, study guides, and hands-on HackAPrompt projects. The AIRTP+ practical exam and one free retake are included. Pass the exam to earn the certification.
Yes. We provide invoices and accept purchase orders. Contact [email protected] for company billing and team purchase options.
You can access the private Discord community to discuss the course with other learners. Recorded lectures, written study guides, and project solutions walkthroughs support independent study.
The exam is a 24-hour practical assessment covering prompt injection, jailbreaking, and AI security defenses. Prepare by reviewing the course materials and practicing in the HackAPrompt playground. One free retake is included.
Basic familiarity with AI tools like ChatGPT is useful. You'll need time to try prompts, read model responses carefully, and document your findings. The course introduces the security concepts used in the exercises.
The course is taught by the team behind HackAPrompt and connects recorded instruction with practice in its playground. You can try attacks, compare your approach with worked solutions, and prepare for the separate AIRTP+ practical exam.
Your course access starts as soon as checkout completes and does not expire. Start the first lecture when you are ready.
Your purchase includes one AIRTP+ exam attempt and one free retake if you do not pass (up to two attempts). The exam is a 24-hour practical assessment covering prompt injection, jailbreaking, and AI security defenses. Follow your exam access instructions to schedule it when you feel ready.
Installment options may be available at checkout, depending on your country and payment-provider eligibility. Checkout shows the available terms before you pay. Companies can also pay by invoice or purchase order.

Ready to Secure Your AI Systems?

Join our AI Security Masterclass and learn the skills you need to protect your organization.

Enroll Now for $1,199

Courses

Resources

Follow Us

© 2026 Learn Prompting. All rights reserved.
Privacy PolicyTerms of Service