Developing a comprehensive strategy for Responsible AI involves establishing governance controls that ensure models are safe, secure, and resilient…

The NIST AI Risk Management Framework (RMF) defines several characteristics of trustworthy AI systems to guide their development and deployment. The core characteristics include ensuring systems are: (1) Reliable and Safe, meaning they perform as intended without causing harm; (2) Secure and Resilient, protecting them from adversarial attacks; and (3) Explainable and Interpretable, allowing stakeholders to understand how the AI model arrives at its decisions. These principles form the foundation for building and managing AI risks responsibly.

D. Proprietary Model Obfuscation is contrary to the principles of explainability and transparency, which are central to building trust in AI systems.

E. Maximum Computational Complexity is a technical attribute, not a characteristic of trustworthiness. In fact, excessive complexity can hinder interpretability and reliability.

1. National Institute of Standards and Technology (NIST). (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0) (Publication No. AI 100-1). U.S. Department of Commerce. Page 6, Section 2.1, "Characteristics of trustworthy AI systems".

2. NIST. (2023). AI RMF 1.0, Page 6, Figure 2, "Trustworthy AI Characteristics". The figure explicitly lists: Safe, Secure and Resilient, Explainable and Interpretable, Privacy-Enhanced, Fair with Harmful Bias Managed, and Accountable and Transparent. Options A, B, and C are direct subsets of these characteristics.