Chapter 6: Explainable Artificial Intelligence and Data Canyons in the Context of Cybernetics

In the last 15 years, the field of AI has witnessed unparalleled advancements, particularly in DL algorithms and LLM (Ahmed et al., 2023; Kasneci et al., 2023; Taye, 2023; Thirunavukarasu et al., 2023; Zhao et al., 2023). These technological breakthroughs present an opportunity in multiple disciplines like healthcare, finance, and education as well as societal governance (Henman, 2020; Kasneci et al., 2023). However, those technological breakthroughs come with their own set of challenges, among which the most prominent are transparency, interpretability, and explainability in AI systems (Gerlings et al., 2020; Gunning & Aha, 2019). This is where XAI comes into play, a promising field dedicated to making AI’s decision-making processes understandable to humans (Barredo Arrieta et al., 2020). XAI is rapidly gaining traction due to its potential to bridge the gap between complex AI systems and human users (Severes et al., 2023). Traditional AI models, often referred to as being black-box, provide little to no insight into how they reach their conclusions (Loyola-González, 2019). The absence of explainability presents an important risk factor, especially where AI systems are integrated into key decision-making processes, where those decision-making processes impact people’s everyday lives. Among the most prominent domains of AI integration in the current age are medical diagnoses, autonomous driving, and financial predictions. Without a clear understanding of how these systems arrive at their decisions, without the reasoning aspect, trust in AI can have severe drawbacks (Ali et al., 2023). The importance of XAI extends beyond technical interest; it has profound practical, social, and ethical ramifications. As AI systems become more pervasive, they play an important role in shaping societal structures and individual lives (Gonçalves, 2001; Shaheen, 2021). In healthcare, for instance, AI-driven diagnostic tools have to be transparent to guarantee that medical professionals can trust and validate recommendations made by those AI systems. In finance, XAI can help moderate risks and further stakeholder trust. In autonomous systems, explainability is crucial for ensuring safety for the whole infrastructure as well as for gaining public acceptance (Atakishiyev et al., 2024). XAI has the potential to fundamentally transform the landscape of cybernetics by enhancing the transparency and interpretability of ML models. Cybernetics, as well as all its underlying parts, can gain a lot from XAI, as it facilitates a greater understanding of complex systems and adds much-needed transparency to decision-making (Anjaria, 2021).