Building Trust in AI-Driven Decision Making for Cyber-Physical Systems (CPS): A Comprehensive Review

AI systems in critical infrastructure require explainability: The paper emphasizes that explainable AI (XAI) is essential for enhancing trustworthiness in cyber-physical systems across sectors like healthcare, agriculture, and autonomous systems, where decisions directly impact physical safety and human welfare.

Transparency and accountability are foundational to trust: The research identifies transparency, accountability, and ethical considerations as necessary pillars for building trust in AI-driven decision-making, suggesting that users and stakeholders need to understand how and why AI systems make decisions in these high-stakes environments.

Bias and interpretation challenges threaten reliability: The paper highlights that complexity in cyber-physical systems creates specific challenges around bias and the interpretability of AI decisions, which can undermine the reliability of automated systems if not properly addressed.

Security and privacy must accompany AI deployment: Beyond algorithmic concerns, the research identifies security and privacy as key challenges that must be managed alongside AI implementation in cyber-physical systems to ensure comprehensive trustworthiness.

Integration of cyber and physical domains amplifies stakes: The seamless integration of digital AI systems with physical infrastructure in CPS means that AI failures or biased decisions have real-world consequences, making responsible AI practices not optional but critical for safe deployment.