Module B: AI Ethics Operationalisation Workflow

Modules

Module B: AI Ethics Operationalisation Workflow

Instructions for:ParticipantTrainer

Related Initiative

AIOLIA

Goal

Module B introduces the AI Ethics Operationalisation Workflow as the central operational framework of the AIOLIA training programme. It translates high-level AI ethics principles into a structured, step-by-step workflow that supports the practical assessment, identification, and mitigation of ethics issues in AI systems across their lifecycle. The module builds on three core building blocks—ethics requirements, practical measures, and ethics tensions—which collectively support the operationalisation of the seven ALTAI requirements in real-world contexts. Illustrated through a practical use case scenario, it guides learners through a 5-step iterative process and establishes the operational backbone of the AIOLIA methodology.

Learning Goals

• Understand the 5-step AI Ethics Operationalisation Workflow and its role in bridging theory and practice
• Use Section 0 of the AIOLIA Issues Identification Checklist to establish the operational context of an AI system
• Identify potential AI ethics risks and map them under the seven ALTAI requirements using Sections 1–7 of the checklist
• Navigate, select, and adapt context-appropriate measures from the AIOLIA Portfolio of Measures (122 technical and organizational measures)
• Recognize and analyze ethics tensions and trade-offs that arise between competing ALTAI requirements
• Perform an Ethics Check to validate findings, ensuring completeness, traceability, and auditable documentation

Duration (hours)

2

For whom is this important?

Doctoral StudentEducator / LecturerEthics reviewersIT / Technical ProfessionalPhD candidatesPolicy Maker / Policy Maker / RegulatorPostdoctoral ResearcherResearch Ethics Committee MemberResearch Ethics ExpertResearcherUndergraduate Student / Graduate Student

Steps

1AI Ethics Operationalisation Workflow

What is this about?

This module introduces the AI Ethics Operationalisation Workflow as the central operational framework of the AIOLIA training programme. It translates high-level AI ethics principles into a structured, step-by-step workflow that supports the practical assessment, identification, and mitigation of ethics issues in AI systems across their lifecycle.

Building on three core building blocks, ethics requirements, practical measures, and ethics tensions, the module demonstrates how the seven ALTAI requirements can be operationalised in real-world contexts. Through a practical use case scenario, learners explore a 5-step iterative process that forms the operational backbone of the AIOLIA methodology.

Why is this important?

While high-level AI ethics principles and guidelines are widely recognized, translating them into concrete engineering choices, organizational policies, and daily practices remains a significant challenge. High-level principles often feel abstract, and real-world implementation inevitably brings competing priorities and ethics tensions.

This module bridges the gap between theory and practice by providing a structured, step-by-step workflow. By learning how to systematically assess risks, apply targeted technical and organizational measures, and navigate ethics trade-offs, learners gain the operational tools needed to design, evaluate, and deploy trustworthy AI systems in real-world environments.

1

AI Ethics Operationalisation Workflow

The AI Ethics Operationalisation Workflow bridges the gap between high-level ethics principles and concrete, day-to-day practices across the AI lifecycle. It is built on three core building blocks: Ethics Requirements (the 7 ALTAI requirements), Practical Measures (a catalogue of 122 technical and organizational safeguards), and Ethics Tensions ("right vs. right" conflicts between competing ethical priorities).

The framework executes through a 5-Step Operational Workflow:

1. Establish Context: Define the system's role, data sources, users, and affected stakeholders using Section 0 of the Issues Identification Checklist.
2. Identify Ethics Issues: Systematically evaluate risks across problem formulation, data engineering, model development, deployment, and monitoring using Sections 1–7 of the checklist.
3. Propose & Prioritise Practical Measures: Browse, assess, adapt, and combine technical measures (for developers) and organizational measures (for management) from the Portfolio of Measures.
4. Identify Ethics Tensions: Analyze and navigate unavoidable trade-offs between competing requirements, such as Privacy vs. Fairness or Technical Robustness vs. Transparency.
5. Perform an Ethics Check: Validate findings and assemble auditable, traceable documentation of all decisions and residual risks.

This workflow applies directly to real-world scenarios, such as the PhiShield case study. When an automated system flags an engineer (Marcus) with an 89% phishing vulnerability score after failing a test during an emergency 14-hour shift, it demonstrates how context-blind evaluations lead to unfair promotion blocks. Resolving such incidents requires applying technical sensitivity analysis (e.g., MS85), establishing a formal Right to Challenge (e.g., MS92), and enforcing mandatory human oversight before taking adverse personnel actions.

AIOLIA Module B (AI Ethics Operationalisation Workflow)

Steps

Other information