Micro-Certification · 90 Minutes

AI Security Essentials

Master the four pillars of AI Governance for regulated industries — discover, test, govern, and observe — in a single focused session. Mapped to ISO 42001, EU AI Act, and NIST AI RMF.

Enroll in the next cohort

Aligned toEU AI ActISO 42001NIST AI RMFCSPAI

Course Duration

90 MIN

Focused, practical learning

Modules

Hands-on, sequential

Certificate

Issued

Upon completion

Course Overview

Practical AI governance — operational from day one.

AI is being deployed across every regulated industry — but governance, security testing, and observability are lagging behind. This 90-minute micro-certification equips IT and Information Security leaders with skills they can apply immediately.

Key objectives

Understand why AI governance is now a regulatory requirement across BFSI, healthcare, and telecom.
Build a complete AI inventory using AI Bill of Materials (AI BOM) methodology.
Conduct AI-specific security testing beyond the OWASP LLM Top 10 using adversarial techniques.
Implement governance controls for AI data, training pipelines, and RAG systems.
Deploy AI observability with token-level tracing for continuous assurance.
Map controls to ISO 42001, EU AI Act, and NIST AI RMF requirements.

Target audience

CISOs & InfoSec Leaders

Leading AI governance and security strategy.

IT Risk & Compliance Teams

Implementing AI controls within regulatory frameworks.

AI / ML Engineers & Architects

Building secure, governable AI applications.

Auditors & GRC Professionals

Assessing AI systems against standards.

DPOs & Legal Teams

Navigating AI-specific privacy and liability.

Product & Business Leaders

Making informed decisions about AI risk.

Prerequisites: Basic information security & risk management knowledge. No AI/ML expertise required.

Course Modules

Four progressive modules, 90 focused minutes.

Each module pairs concepts with a hands-on exercise on a banking-chatbot case study, so the learning maps cleanly to real production systems.

Module 01 · Discover & Inventory

AI Bill of Materials (AI BOM)

~20 min

What AI BOM is and why it matters for governance.
Build a structured inventory of all AI components.
Identify models, versions, owners, data, dependencies, and risks.
Categorise internal AI, third-party AI, SaaS AI, and shadow AI.
Hands-on: create an AI BOM for a banking chatbot case study.

Module 02 · Test & Validate

AI Red Teaming & Security Testing

~25 min

AI-specific vulnerabilities: prompt injection, data extraction, adversarial attacks.
Move beyond the OWASP LLM Top 10 to technique-based testing.
Build threat profiles for each AI inference point.
Business-aware testing: map workflows, define misuse scenarios, mutate attacks.
Hands-on: create threat profiles for the banking chatbot.

Module 03 · Govern Data Lifecycle

AI Data Governance

~20 min

Why data is the highest-risk factor in AI systems.
Required controls: PII scanning, bias detection, quality validation.
Data governance for fine-tuned LLMs, RAG systems, and ML decisions.
Mitigation controls: LLM firewalls, guardrails, access restrictions.
Hands-on: apply data controls to the chatbot RAG pipeline.

Module 04 · Monitor Continuously

AI Observability

~25 min

Why traditional SIEM and logging fail for AI systems.
Three monitoring layers: performance, security, and drift.
Token-level tracing with OpenTelemetry.
Map observability to EU AI Act, ISO 42001, and BFSI mandates.
Hands-on: design an observability plan for the chatbot.

Skills Gained

What you'll be able to do when you finish.

Build and maintain an AI Bill of Materials (AI BOM).
Conduct AI-specific red teaming and adversarial testing.
Implement data governance controls for AI training and RAG pipelines.
Deploy AI observability with token-level tracing.
Map AI controls to ISO 42001, EU AI Act, and NIST AI RMF.
Create threat profiles for AI inference points.
Evaluate third-party and SaaS AI vendor governance.

Your Certification Path

From Essentials to practitioner.

AI Security Essentials is the entry point. From here, you can advance to CSPAI and then operate the full PRISM platform.

You are here