EC-Council Certified Offensive AI Security Professional (C|OASP)

Price
Net €
VAT €

Price
Price on Request

Duration
5 days

For companies and job seekers:
this course is 100% fundable!

Location

Course Language
English

Training Solutions
Online Live

Request a consultation Add to favorites Print as PDF

Attacks are becoming more sophisticated—and increasingly rely on AI. What is needed is a deep understanding of how offensive techniques are combined with artificial intelligence to test systems under realistic conditions and identify vulnerabilities early on.

Key Topics

Adversarial AI and attack models
Prompt injection and model exploits
Automated vulnerability analysis
Red teaming with AI support
Securing ML and LLM Systems

Prerequisites
Fundamentals of IT security and networking, and some familiarity with machine learning

Target Audience
Security professionals, penetration testers, red teams, IT architects

Relevant expertise for a security landscape where traditional methods are no longer sufficient and intelligent attack simulation is becoming the norm.

Print as PDF

Course Content

Offensive AI and Methods for Hacking AI Systems

Understand the fundamentals of AI and machine learning from an offensive security perspective.
Identify attack surfaces, threat landscapes, and attacker techniques in the field of AI, aligned with MITRE ATLAS.
Apply methods, frameworks, and risk implications when hacking AI systems.
Classify taxonomies and models for AI attacks.
Define the fundamentals of offensive AI and the fundamentals of securing AI systems.
Provide an overview and mapping of the OWASP LLM & ML Top 10 (2025) to AI threats and governance considerations.

AI Awareness and Vulnerability Mapping

Use of OSINT tools and techniques to identify and profile AI resources.
Gathering information from AI data sources and training pipelines.
Identify and map AI attack surfaces using publicly available information.
Identify AI endpoints, services, APIs, and exposed parameters.
Identify and analyze AI models and vector stores from an attacker’s perspective.
Assess OSINT exposure and apply security measures to mitigate risks.
Leverage AI threat intelligence to support continuous monitoring and defense readiness.

Vulnerability scans and fuzzing in the field of AI

Understand the fundamental principles of vulnerability analysis and threat detection in AI.
Use tools and techniques to scan for vulnerabilities in AI models, pipelines, and deployments.
Apply practical fuzzing methods specifically tailored to AI systems and model interfaces.
Integrate scanning and fuzzing into AI security workflows to proactively mitigate risks.

Prompt Injection and Attacks on LLM Applications

LLM architecture, trust boundaries, and associated attack vectors.
Application of prompt injection and jailbreaking techniques in real-world LLM applications.
Identification of risks related to the disclosure of sensitive information and the loss of system prompts.
Assessment of vulnerabilities arising from improper output handling and threats posed by misinformation.
Application of advanced prompt-based attack techniques and exploit strategies.
Implementation of principles for secure LLM application design and defensive control measures.

Adversarial Machine Learning and Model Privacy Attacks

Identify core categories of adversarial machine learning attacks.
Perform practical adversarial input attacks across various data modalities.
Apply attack techniques in the areas of privacy, inference, and model extraction.
Evaluate methods for assessing robustness, trustworthiness, and risks.
Implement defense strategies for data privacy and model resilience.

Attacks on data and training pipelines

Understand the architecture of AI data and training pipelines, as well as their vulnerabilities.
Implement practical techniques for data corruption and attack scenarios.
Apply the insertion of backdoors and Trojans during model training.
Implement security measures to protect data and training pipelines.

Action-Oriented AI and Model-to-Model Attacks

Understand the architecture and attack surface of action-oriented AI.
Apply techniques to exploit excessive agency and autonomy.
Identify attack vectors across multiple LLMs as well as model-to-model attack vectors.
Assess risks related to “denial-of-wallet” and unlimited resource consumption.
Conduct attacks on AI workflows and orchestration layers.
Implement defense strategies to secure action-oriented AI applications.

Attacks on AI infrastructure and supply chains

Understand AI infrastructure components and system integration architectures.
Identify vulnerabilities in AI systems, frameworks, and deployment pipelines.
Analyze the misuse of tools, plugins, and APIs in AI-powered applications.
Assess threats to the AI supply chain and dependency risks (in-depth analysis).
Implement strategies to secure AI infrastructure and supply chains.

Security testing, evaluation, and safeguarding of AI

Understand security testing methods and evaluation techniques in the field of AI.
Apply red team frameworks for the offensive assessment of AI.
Identify, validate, and document security vulnerabilities and risks related to AI.
Implement best practices for securing and mitigating risks in AI systems.

AI Incident Response and Forensics

Detect and respond to AI-related security incidents.
Collect and analyze AI logs, telemetry data, and digital evidence.
Determine the causes as part of the post-incident analysis.

Your training will be 100% funded

Our funding schemes are available exclusively to participants who live or work in Germany. Funding is not available outside Germany.

For career starters, job seekers, career changers, startups, experienced professionals, those looking to advance their careers, specialists, etc.

The QCG program for everyone in the company

Further information

The education voucher for all job seekers

Further information

Frequently Asked Questions

What are the specific benefits of C|OASP certification?

Practical expertise in targeting and securing AI systems—exactly the skills that companies are urgently seeking right now.

Why is offensive AI security in such high demand right now?

Because AI models create new vulnerabilities—those who understand them can identify real security gaps before others do.

What risks arise without expertise in AI security?

Manipulated models, data breaches, and undetected vulnerabilities—often resulting in massive financial and reputational damage.

What sets C|OASP apart from traditional security certificates?

This isn't about theory, but about active attacks on AI systems—a clear advantage over standard certifications.

What specific skills are taught?

Prompt injection, model exploitation, adversarial attacks, and securing AI systems—ready to apply on the job.

How quickly can this knowledge be put into practice?

Right away – the content is designed to address real-world threats and can be applied directly to projects or within the company.

Why invest in AI security now?

Because the market is evolving rapidly, those who build up expertise early on will secure clear career and salary advantages.

Do you have any further questions? Please contact us.

Select your nearest center: *

0800 258 258 055

website@newhorizons.de

First Name *

Surname *

Telephone Number *

Postcode *

E-Mail *

I would like to receive funding for further training – I am currently: *

Company Name

Your message

I hereby confirm that I have taken note of the privacy notice. *