Table 2 Framework Pillars: Mapping Foundational Principles for SMEs

Data

Integrity, authenticity, provenance, and compliance

Ensures data integrity through bias detection, anonymisation, and compliance with privacy regulations. Provenance tracking safeguards against poisoning, skewness, and regulatory breaches, building trust in downstream decision-making.

Algorithms

Fairness, robustness, accountability, and resilience

Serve as computational engines that transform data into actionable intelligence. Embedding ethical parameters and explainability mechanisms in algorithm design helps SMEs avoid bias, prevent discriminatory outcomes, and enhance stakeholder confidence.

Human Oversight

Ethical anchoring, contextual awareness, shared accountability

Human-in-the-loop, on-the-loop systems ensure domain expertise as well as ethical reflection in high-stakes decisions, with accountability shared between humans and AI systems.

Model Architecture

Secure-by-design, transparency, alignment with context

Provides technical scaffolding. Secure-by-design principles enhance resilience, transparency and explainability supports stakeholders interpretation, and context-aware architectures align outputs with SME-specific goals and requirements.