Hacking AI Systems: Exploits, Defenses, and the Future of Red Teaming

Agenda for the Masterclass

Introduction
- Why AI systems are a new attack surface
- Real-world cases of AI misuse and vulnerabilities
- Setting expectations for demos and outcomes
Understanding the AI Attack Surface
- Components: prompts, embeddings, plugins, external tools, memory
- Differences between traditional app pentesting vs AI pentesting
- Mapping threats to frameworks (MITRE ATT&CK for AI, OWASP LLM Top 10)
Hands-On Demo: Exploiting AI Systems
- Prompt Injection → bypassing guardrails
- Sensitive Information Disclosure → unintended data leakage
- System Prompt Leakage → extracting hidden instructions
- Excessive Agency → agents executing unintended actions
Pentesting Methodology for AI Systems
- Reconnaissance: AI-powered OSINT, banner grabbing, fingerprinting
- Exploitation: adversarial prompts, poisoning embeddings
- Post-exploitation: persistence in agent workflows, unauthorized tool use
- Safe lab setup for AI pentesting (VMs, sandboxed APIs)
Defensive Countermeasures (Blue Team View)
- Guardrails: input validation, adversarial testing, monitoring
- Human-in-the-loop strategies
- Logging and anomaly detection for AI interactions
- Case study: spotting prompt injection attempts
Interactive Q&A and Scenario Discussion

Why Attend This Masterclass

Get CPE Certificate

FREE Career Guidance & Mentorship

Learn from Industry Experts

Ashish Dhyani

10+ Years of Experience

Upcoming Events for You

FREE Training

AI-Powered Cyber Defense: The Future of Threat Detection & Response

30 Jun (Tue) , 08:00 – 10:00 PM (IST)

294People have registered

FREE Training

Practical AI Security Masterclass

02 Jul (Thu) , 08:00 – 10:00 PM (IST)

267People have registered

FREE Training

Become an AI Security Leader: Why AAISM Matters in the AI Era

03 Jul (Fri) , 08:00 – 09:00 PM (IST)

230People have registered

Paid Masterclass

AI-powered Security Operations Bootcamp How LLMs Are Exploited by Hackers

11 – 12 Jul (Sat – Sun) , 10:00 AM – 02:00 PM (IST)

Save a seat before the spots fill out!

<ul>
<li>Introduction
<ul>
<li>Why AI systems are a new attack surface</li>
<li>Real-world cases of AI misuse and vulnerabilities</li>
<li>Setting expectations for demos and outcomes</li>
</ul>
</li>
<li>Understanding the AI Attack Surface
<ul>
<li>Components: prompts, embeddings, plugins, external tools, memory</li>
<li>Differences between traditional app pentesting vs AI pentesting</li>
<li>Mapping threats to frameworks (MITRE ATT&CK for AI, OWASP LLM Top 10)</li>
</ul>
</li>
<li>Hands-On Demo: Exploiting AI Systems
<ul>
<li>Prompt Injection → bypassing guardrails</li>
<li>Sensitive Information Disclosure → unintended data leakage</li>
<li>System Prompt Leakage → extracting hidden instructions</li>
<li>Excessive Agency → agents executing unintended actions</li>
</ul>
</li>
<li>Pentesting Methodology for AI Systems
<ul>
<li>Reconnaissance: AI-powered OSINT, banner grabbing, fingerprinting</li>
<li>Exploitation: adversarial prompts, poisoning embeddings</li>
<li>Post-exploitation: persistence in agent workflows, unauthorized tool use</li>
<li>Safe lab setup for AI pentesting (VMs, sandboxed APIs)</li>
</ul>
</li>
<li>Defensive Countermeasures (Blue Team View)
<ul>
<li>Guardrails: input validation, adversarial testing, monitoring</li>
<li>Human-in-the-loop strategies</li>
<li>Logging and anomaly detection for AI interactions</li>
<li>Case study: spotting prompt injection attempts</li>
</ul>
</li>
<li>Interactive Q&A and Scenario Discussion</li>
</ul>

By submitting your contact details, you acknowledge and agree to our Terms of Use and Privacy Policy.

I agree to receive your marketing and promotional emails

201 People have registered