When an AI decision is challenged, can your system justify it?

In production, an AI decision depends on more than a model. Workflows, configuration, human reviews, and other services all affect the outcome. Each component can work as intended, yet the final decision can still be hard to explain.Foundations for Keeping AI Responsible approaches this as a software design problem. It introduces the RAIMO framework and the AI Passport as ways to make responsibility visible in the design and in the decisions the system makes.Using an AI-assisted lending workflow, the book explores what it takes to know which rules applied, who was responsible, and whether a decision was allowed to go ahead.

A Systems Approach to Accountable Decisions in Production

Foundations for Keeping AI Responsible

A Systems Approach to Accountable Decisions In production

Table Of Content

Part I — When Responsibility Breaks

  • Chapter 1: Why Responsibility in AI Is a Systems Problem

  • Chapter 2: What Changes When Responsibility Is Enforced

  • Reference: Mapping Responsible AI Principles to the Five Pillars

  • Reference: Mapping the Five pillars to the EU AI Act

  • Interlude: What Meaningful Human Oversight Requires

Part II — Designing for Responsibility

  • Chapter 3: Why Responsible AI needs to be designed

  • Reference: Mapping the Five Pillars to System Capabilities

  • Interlude: Different Frameworks, Different Jobs

  • Reference: What Does a Good Decision Record Look Like?

  • Interlude: How Requests Flow Through the System

  • The Next Five Chapters

  • Chapter 4: Defining Responsibility: Context boundaries

  • Chapter 5: Binding Responsibility: Runtime Ownership

  • Interlude: The cost of enforcement

  • Chapter 6: Preserving Responsibility: Runtime Integrity

  • Chapter 7: Controlling Responsibility: Execution contraints

  • Chapter 8: Inspecting Responsibility: Decision Observability

Part III — System Behavior

  • Chapter 9: Responsibility In A Multi-decision Workflow

  • Closing note

  • AI Passport and decision record templates

  • Reference: Reviewing Responsibility in System Design

  • Reference: EU AI Act Role Triage for Engineering Teams

  • Reference: Reviewing EU AI Act Risk in System Design

  • Reference: Who Owns What at a Decision Point