Web attacks aimed at deceiving Large Language Models (LLMs) are primarily known as Prompt Injection.
These attacks involve introducing malicious natural language instructions to manipulate the behavior of the Artificial Intelligence model so that it performs unwanted actions, bypasses security restrictions, or leaks confidential information.
Internet Security Auditors analysts will develop some Portswigger labs to explain certain vulnerabilities related to Prompt injection.
During this webinar we will cover the following topics:
- What is an LLM?
- Types of attacks on LLMs
- Prompt Injection
- Indirect Prompt Injection
- Training data poisoning
- Sensitive training data leakage
- AI-Powered scanners
- Protection against attacks on LLMs
Why should you attend?
LLMs are here to stay, and attacks against them too. If you work with AI, development, or cybersecurity, this session will help you understand the real risks of LLMs.
Through practical labs, we will see how the most common attacks occur and how to defend artificial intelligence-based applications.
When
October 29, 2026
From 9:00 to 9:45 (MX time) | 10:00 to 10:45 (BO time) |
11:00 to 11:45 (CH time) | 16:00 to 16:45 (ES time)
Speaker
Mario Valiente Catalán
Security Auditor
eJPT, OSCP, CRTO, BSCP, CRTP,
Windows Evasion Specialist,
Malware Development Specialist
Roger Martín Campanario
Security Auditor
OSCP, CARTP, CARTE
🚀 Secure your spot and participate in a session that will allow you to deeply understand the new standard and its real impact on your organization.
Completely free webinar
Register now!