Glossary
R
17 entries beginning with R, the same definitions the articles use.
RACI matrix
A responsibility-mapping tool identifying who is Responsible, Accountable, Consulted and Informed for each AI and ML application, used here to capture stakeholders and survey their requirements for the register.
Related: The Board in the machine
Reasoned Indicator Concept
A fourth indicator type alongside lagging, leading, and predictive indicators; where those estimate or forecast, a reasoned indicator proves what is possible, impossible, or must hold true under any combination of inputs.
First introduced in: From Probable to Provable: What Automated Reasoning Means for the Board
Related: From Probable to Provable: What Automated Reasoning Means for the Board
See also:Predictive indicatorMaximum Fidelity
Reasoning Gap Concept
The gap between the four legal safeguards required for solely automated decisions and a system’s actual ability to interrogate and explain its own decisions, a capability built into rule-based systems but absent by default in probabilistic ones.
First introduced in: The Reasoning Gap: The Capability the Law Now Demands of Boards
See also:rule-based system
Reasoning models
AI models that perform more computation at the response layer to reason through problems, consuming far more inference per query than predecessors because the work happens after the prompt.
Related: The Headroom Argument: Why AI Efficiency Means More Compute, Not Less
See also:Inference
Red-teaming
Deliberately stress-testing an AI system by trying to make it fail or misbehave, to surface vulnerabilities before deployment.
See also:AI safetyAI vulnerability management
Redeployment Dividend Asset
The strategic value released when the hours AI frees are redirected to differentiated work rather than eliminated; the headline line of any remaking’s business case. It is realised only by deliberate redeployment: token retraining programmes forfeit it, merely delaying displacement and undermining Board confidence.
See also:Reinvention DividendSelective AtrophyUndifferentiated work
Regulatory arbitrage
A strategy of conducting research and development under permissive US frameworks while building trust through EU compliance for market entry, exploiting differences between divergent regulatory regimes.
Related: Why Boards Need to Watch the EU's General-Purpose AI Code of Practice
See also:EU AI ActPrincipled Standardisation
Reinforcement Learning
A machine learning approach in which algorithms learn through trial and error, used in dynamic pricing, personalised recommendations, and autonomous systems that continuously learn from interaction.
Related: Harnessing AI for organisational change led from the Board
See also:Machine learning
Reinvention Dividend Concept
The compounding return organisations gain by sustaining a flywheel of low-cost, fast experimentation and change; even failed experiments yield insights that inform subsequent tactics and decisions about what happens next.
First introduced in: Why now is not the time to take your foot off the gas
See also:Redeployment Dividend
Remake Concept
Every organisation does four kinds of work: thinking, deciding, creating, and delivering, and AI is remaking all four. Remake is how a Board governs that remaking and how an organisation executes it: establishing what the work is, who does the remaking, who is accountable for it (a gate, not a description; where genuine agency or durable accountability is absent, the remaking does not begin), what the verdict on the work is (Stop, Keep, or Change), and how it is done through the ADAPT motion. The mature successor to the Complete AI Adoption Framework.
First introduced in: Remake
See also:ADAPTStop Keep Change ModelThe Great RemakingComplete AI Adoption Framework
Report on Compliance
The independently issued report setting out a Qualified Security Assessor’s view of PCI DSS compliance, covering not just shopping-cart software but the physical hosting environment and servers where a website runs.
Responsible AI
The practice of designing, deploying and governing AI in ways that are ethical, fair, transparent and accountable.
Related: Demystifying AI: My Chief Wine Officer Talk
See also:EU AI ActAI safetymodel card
Retrieval
Grounding a model’s outputs in an organisation’s own documents and data; the right tool for accuracy, and the wrong tool for changing how a model weighs values.
Related: Agentic AI: Strip Away the Hype and Understand the Real Strategic Choice
Risk-based triage
The rapid first-week sorting of disclosed shadow AI discoveries by risk level using a red/amber/green system, prioritising high-risk items for immediate action while identifying quick wins across departments.
Related: After the AI Amnesty: Practical Steps to Operationalise Discovered Shadow AI
See also:Shadow AIAI Amnesty
RLHF
Reinforcement learning from human feedback, a method providers use to align a model’s behaviour to human preferences, one source of the value dispositions a model carries into deployment.
See also:Reinforcement Learning
Robotic Process Automation
Rule-based software automation that executes defined, repeatable tasks deterministically; still the primary orchestrator of many mission-critical business flows despite the generative AI surge.
Related: Introducing the AI Stages of Adoption: A framework for understanding AI readiness in your business
See also:Generative AI
rule-based system
A system whose decision logic (eligibility rules, scoring thresholds) is explicit and inspectable, so the reasoning behind any individual decision can be shown directly, making legal safeguards straightforward to deliver.
Related: The Return of Traditional AI: Organisations Are Rethinking Their LLM-First Strategies
No R entries match. or try the site search.