[A] Accuracy and Consistency
Focuses on correctness and alignment across information, processes, and expectations. This dimension ensures that facts, data, concepts, terms, and procedures are consistent, historically validated, and free from contradictions. It also includes individuals’ behaviors, beliefs, biases, and needs, as well as conversations, plans, and work products. Focusing on accuracy and consistency supports behaviors that minimize harm and maintain reliable operations.
[C] Certainty of Course Corrections
Highlights the importance of monitoring, feedback, and intervention to reduce risk. It covers self-correction at the task level as well as correction by an intervenor, whether onsite or remote. By enabling continuous adjustment, personalization, and supervision, this dimension ensures that deviations are detected and corrected quickly, maintaining low-risk outcomes and predictable performance.
[I] Certainty of Interpretation
Addresses the accuracy and reliability of understanding by both humans and AI. It includes correctly interpreting priorities, values, and contextual nuances, as well as accounting for cultural, ethical, regulatory, and geographical differences. By focusing on correct interpretation, organizations can ensure that outputs align with expectations and that risk from miscommunication or misunderstanding is minimized.
[D] Certainty of Delegation
Focuses on safely assigning tasks to AI systems or individuals while maintaining quality and control. Tasks with low variance can be delegated with minimal instruction and reviewed either onsite or remotely. However, tasks requiring differentiation, expertise, or nuanced judgment, such as creative or specialized work, require supervision or cannot be delegated. By clarifying what can and cannot be delegated and the expected quality of outputs, this dimension reduces errors and maintains consistent performance.
[P] Certainty of Position, Pattern, and Path
Emphasizes predictability and stability in spatial, procedural, and temporal systems. It covers the position and state of objects, observable patterns, anomalies, and the flow of activities and processes. By maintaining clarity around relationships, timing, and sequences, organizations reduce uncertainty and risk while improving coordination, planning, and execution.
[L] Risks from Learnability
Captures the ease of adoption and training for both humans and AI. It considers the transparency of automation, the support for augmented processes, and the ability for users to self-learn or follow guidelines. By designing systems that are easy to understand and progressively augmentable, this dimension minimizes risks associated with misuse, poor adoption, or training gaps.
[A] Certainty of Analysis
Addresses the reliability of reasoning, problem-solving, and decision-making. Even when initial answers are partial or incorrect, structured analysis allows organizations to converge on the correct solution. This dimension reduces risks associated with misjudgments and supports behaviors that prevent harm or errors.
[N] Certainty of No Intervention
Focuses on tasks and processes that can safely proceed without oversight or correction. Low-variance, low-risk tasks can be executed repeatedly without supervision, while high-variance or high-risk tasks still require review and intervention. This dimension ensures that repeated operations, simulations, or data processing do not compound hidden errors and remain reliable over time.
