Training Module
AI Fundamentals II
AI uncertainty, limitations & common failure modes across predictive and generative AI systems
Training module overview
AI systems operate under uncertainty. Their behaviour shifts with data, context, integration choices, and human interaction. When these limits are not understood, organisations either over-trust or over-block AI, creating governance and operational risk.
This module develops a clear, structured view of AI limitations and failure modes across predictive and generative systems. Participants learn where uncertainty originates, how failures emerge in real environments, and how to interpret AI outputs and technical evidence with appropriate judgment.
Applicable environments
This module applies to organisations for which artificial intelligence is relevant. It supports professionals who need a solid understanding of AI-related concepts, terminology, and context.
Target audience
AI management system managers and implementers working with technical teams
Governance, risk, and compliance professionals who need AI domain fluency
Product owners and process owners responsible for AI-enabled services
Auditors who need a shared baseline understanding of AI systems (not audit craft)
Anyone who wants to get a basic understanding of AI fundamentals
Decision support
Is this module for you?
It is a good fit if you…
want to understand why AI outputs are signals, not facts.
need a realistic mental model of uncertainty in predictive and generative AI.
want to recognise common AI failure modes in real operational contexts.
need to reason about reliability, limits, and confidence without false certainty.
want to interpret AI behaviour without relying on vendor claims or tooling.
If most of the points above apply, this module is likely a good fit.
It may not be the best fit if you…
are looking for AI risk assessment methods or control design.
expect statistical deep dives or model-level optimisation techniques.
want implementation playbooks, monitoring setups, or lifecycle processes.
already have advanced AI research or data science expertise.
Agenda
Why limitations and uncertainty are central to AI governance
Where uncertainty comes from in AI systems
Model behaviour limits in predictive ML
Model behaviour limits in generative AI
Data-related failure modes
System and socio-technical failure modes
Case-based workshop
Show detailed agenda...
Learning outcomes
Key outcomes
Explain the main sources of uncertainty in AI systems and how they affect outcomes
Recognise common failure modes in predictive machine‑learning models and generative AI systems
Describe socio‑technical failure modes where human and organisational factors interact with AI
Additional capabilities
Identify data pipeline issues that contribute to model drift and bias
Conduct structured failure‑mode walkthroughs to anticipate how AI performance may degrade
Communicate limitations and uncertainty to stakeholders to support responsible AI use
Additional benefits
Learning materials
Slide deck
Participant workbook
Templates & tools
Practical, reusable artefacts to apply the module directly to your organisation.
AI uncertainty map (sources and observable signals)
Failure mode catalogue (predictive, generative, and system-level)
Case walkthrough canvas (data → model → integration → use)
Evidence question set for technical walkthroughs
Confirmation
Certificate of completion
Module ID
HAM-AI-DF-02
Discipline
Domain
Audience
Auditor
Manager
Executive
Language
English
Delivery
Live virtual
Duration
7 h
List price
CHF 550
Excl. VAT. VAT may apply depending on customer location and status.
Delivery & learning format
Virtual live teaching
This module is delivered live, with a strong focus on discussion, practical application, and direct interaction with the instructor.
Sessions work through realistic examples, clarify concepts in context, and apply methods directly to participants’ organisational realities.
Custom delivery options
For organisations with specific constraints or learning objectives, the module can be adapted in format or scope, including in-house delivery and contextualised case material.
For an optimal learning experience
Preparation guidance
This module is designed as part of a modular training approach. Topics are deliberately distributed across modules and are not repeated in full, in order to avoid unnecessary redundancy. Each module is self-contained and can be taken on its own. Where prior knowledge or experience is helpful, this is indicated below so you can decide whether any preparation is useful for you.
Assumed background
This module assumes baseline familiarity with core AI concepts and system types (data, training vs. inference, and common AI architecture patterns). Participants should also be comfortable reading high-level technical descriptions (services, APIs, data stores).
Helpful background includes:
Basic understanding of digital services and dependencies (applications, interfaces, data flows)
Familiarity with common IT control concepts (access control, logging, encryption) at a conceptual level
Preparatory modules
Supporting modules (optional)
Helpful if you want to deepen related skills, but not required to participate effectively.
