This workbook is a facilitator’s guide to delivering AI ethics training across public institutions. It covers AI fundamentals, public sector use cases, and governance models—paired with activities grounded in UK government experience and policy frameworks.
📘 What’s Covered
The workbook is the first in an eight-part series under the AI Ethics and Governance in Practice Programme, funded by UKRI and designed by The Alan Turing Institute’s Public Policy Programme. Its focus is capacity-building, not policy drafting — and that’s a strength.
This opening module is built around two main sections:
1. Key Concepts
It demystifies AI for non-technical professionals by offering plain-language definitions and structured walkthroughs:
- What AI and ML systems are, technically and socially
- The AI/ML project lifecycle, from problem framing to post-deployment
- Types of machine learning (supervised, unsupervised, reinforcement)
- Common public sector use cases (from predictive policing to energy grid management)
- The role of ethics in AI, rooted in accountability, safety, and fairness
The workbook introduces key frameworks such as:
- CARE & ACT principles for reflective governance
- SSAFE-D (Sustainability, Safety, Accountability, Fairness, Explainability, Data stewardship)
- The SUM Values for ethics-first delivery culture
2. Activities
The second half includes workshop-ready exercises to translate ideas into practice. These are designed for use by in-house facilitators (called AI Ethics Champions), and include:
- Creating a team’s “collective image” of AI
- Mapping the project lifecycle
- Role-playing ethical dilemmas in AI deployment
- Applying CARE/ACT to mock or real use cases
The programme also includes domain-specific workbooks (e.g. AI in social care, policing, education, urban planning), making the training highly contextual.
💡 Why it matters?
Too many AI governance efforts stop at frameworks or policy docs. This one rolls up its sleeves. By equipping civil servants with language, tools, and exercises they can run themselves, the workbook shifts AI ethics from theory to organisational muscle memory. It doesn’t assume everyone is a data scientist — it assumes everyone can (and should) ask good governance questions.
🔍 What’s Missing
The workbook is light on cross-border interoperability — there’s little discussion of how these UK-based frameworks might align with EU AI Act requirements, ISO standards, or NIST practices. Also, while it offers tools for values-based governance, it’s less prescriptive about risk thresholds, audits, or incident response — likely covered in later modules, but still notable.
🎯 Best For
Public sector staff, especially those in policy, service delivery, or digital transformation roles. Also valuable for ethics committees, AI governance officers, and facilitators running AI literacy programs. Great fit for teams implementing responsible AI mandates in complex bureaucracies.
📚 Source Details
- Title: AI Ethics and Governance in Practice: An Introduction (Facilitator Workbook)
- Authors: David Leslie et al., The Alan Turing Institute
- Date: 2023
- Funding: UKRI, EPSRC
- Use case: Workshop facilitation, public sector training
- Frameworks used: CARE & ACT, SSAFE-D, SUM Values
- Available via: turing.ac.uk/ai-ethics-governance
