TRAINING

AI Systems Security

This training gives cybersecurity and AI professionals practical skills to secure AI systems, end to end. Through live sessions and lab simulations, you will learn to identify AI threats, design secure architectures, and defend against real-world AI attacks.

Start dates

12 November 2026

Duration

24 hours instructor-led and 24 hours self-paced training over 6 weeks

Course fee

CHF 1’990

Level

Expert

Language

English

Certificate

AI Systems Security Expert

BROCHURE DOWNLOAD

Events

Testimonials

The trainign offered valuable technical insights into model poisoning and other threats. My favorite part was the hands-on challenge of cracking Level 8 of the Gandalf LLM test; it was a brilliant way to practice the adversarial mindset required for defense. I highly recommend this course to developers and security engineers.

Jamie Mcallister

Former student

The training provided me with practical skills and I am already applying them in my work. What I particularly valued was the strong practical focus and the opportunity to learn from a recognized institution like the Swiss Cyber Institute. I would definitely recommend this training to anyone looking to strengthen their profile and stay up to date in this fast-moving field.

David Alvarado

Former student

A truly professional institute with well-trained instructors, dedicated mentors, and friendly staff; expert-level training in cybersecurity and AI that’s practical, well-structured, and immediately applicable.

Patrik Birchmeier

Former student

The team at the Cyber Security Institute is great. Firstly, they are always available for guidance and secondly, they are very professional in what they do.

Tomislav Pahlina

Former student

I took a course and was very satisfied with both the content and the excellent organization by the company.

Simone Fossati

Former student

Great place to study. Well-organized, creating a prolific learning environment. Well-prepared lecturers with a vast knowledge in their domain.

Yvonne

Former student

Great team! Very supportive in your studies path, the learning is very good also with a plataform that makes things more organized.

Rúdi Machado Delgado

Former student

Who is this course for?

Prerequisites

Participants should understand core security concepts, basic machine learning workflows, and risk management. Experience with Linux, Python, networks, and cloud environments is recommended for hands-on exercises.

Target audience

The course is ideal for security professionals, ML developers, and data scientists who want to understand and defend against AI-specific threats. It also suits SOC analysts and architects aiming to design and secure AI systems across their lifecycle.

Foundations and expanding attack surfaces

Learn how modern AI systems introduce new ways to attack, from prompt injection and data poisoning to stolen model weights and cross-modal exploits. You will map these threats to real architectures and practice defending a sample AI service in a hands-on lab.

AI threat modelling and secure design

Use AI-specific threat frameworks to identify risks from both human and agentic attackers. You will assess compliance readiness and model threats for supervised vs. autonomous AI systems in a collaborative exercise.

Secure AI design and DevSecMLOps

Design and deploy AI systems with built-in security, from model signing and encrypted storage to supply-chain protections and CI/CD security gates. Then apply it all in a hands-on lab using open-source tools.

Adversarial attacks and defensive engineering

Learn to run and automate adversarial attacks such as FGSM, PGD, data poisoning, and Trojan inputs. You will also test cross-modal exploits and measure model robustness on benchmark platforms. In a red-team/blue-team lab, you will apply defenses like adversarial training and certified hardening.

Privacy, fairness, and ethical risk management

Protect data using methods like differential privacy and federated learning. Learn auditing AI systems for bias with fairness metrics and creating transparency reports for governance. Through case studies and group discussions, you will also apply ethical frameworks to real-world challenges.

Incident response, forensics, and AI resilience

Learn to monitor and detect AI-specific threats and contain incidents such as prompt leakage. Collect forensic evidence and practice managing a full breach scenario in a collaborative team exercise.

Learning objectives and competencies

This course teaches you to secure AI systems across their entire lifecycle, from data protection to incident response. You will learn to design secure architectures, defend against AI-specific attacks, and apply frameworks that ensure privacy, fairness, and compliance at every step.