Catalog of General Ethical Requirements for AI Certification

Six core ethical principles should guide AI development: The paper identifies fairness, privacy and data protection, safety and robustness, sustainability, transparency and explainability, and truthfulness as essential pillars for trustworthy AI, providing a structured framework that developers can use to evaluate their systems.

Ethical requirements need practical implementation tools: Rather than abstract ideals, the paper offers value-specific recommendations and concrete criteria for risk assessment, helping organizations translate ethical principles into actual technical and operational practices.

AI systems should be categorized by risk level: The guidance aligns with the EU AI Act's approach of categorizing AI applications based on their potential harms, suggesting that different levels of ethical scrutiny and certification requirements should apply depending on the risks posed by specific AI systems.

Certification frameworks can establish minimum ethical standards: By presenting this work as a "potential blueprint" for AI certification, the paper suggests that standardized ethical requirements and assessment processes could help ensure baseline trustworthiness across the AI industry.

Sustainability deserves equal weight with traditional AI ethics concerns: The inclusion of sustainability alongside fairness, privacy, and safety indicates that responsible AI development must consider environmental and resource impacts, not just algorithmic bias or data protection.