The rapid growth of e-commerce and urbanization has intensified the complexity of last-mile logistics, creating significant environmental and operational challenges. This chapter examines the transformative role of technology innovations and digital solutions in achieving sustainable last-mile delivery systems. It explores cutting-edge technologies including autonomous vehicles, drones, electric fleets, blockchain for supply chain transparency, artificial intelligence for route optimization, Internet of Things (IoT) sensors, and crowdsourcing platforms. The chapter analyzes how these innovations address traditional challenges such as traffic congestion, carbon emissions, delivery failures, and cost inefficiencies.
Explainability and transparency have emerged as foundational pillars in the secure deployment of artificial intelligence (AI) systems, especially large language models (LLMs). This chapter examines the evolving landscape of explainable AI (XAI) architectures through the lens of cybersecurity, adversarial robustness, and regulatory compliance. The authors survey core XAI methodologies—including LIME, SHAP, mechanistic interpretability, attention attribution, and causal tracing—evaluating their effectiveness against adversarial threats such as jailbreaking, prompt injection, data poisoning, and hallucination exploitation. The dual nature of XAI is critically examined: while transparency mechanisms bolster defense and trust, they simultaneously introduce novel attack surfaces that adversaries can exploit to subvert explanation systems.
Manufacturing industries account for a significant share of global energy consumption and greenhouse gas emissions, making resource efficiency an urgent priority. This chapter examines the integration of Geographic Information Systems (GIS) with Industrial Internet of Things (IIoT) technologies, digital twins, and artificial intelligence to enable spatial energy mapping and resource efficiency in manufacturing environments. Energy consumption data collected from smart submeters and edge-computing platforms can be visualized as geospatial layers to identify energy hotspots, quantify waste sources, and prioritize intervention strategies.
Hyperspectral imaging (HSI) enables real-time, non-invasive tissue characterization by capturing biochemical signatures across hundreds of spectral bands simultaneously. However, the computational demands of processing vast hyperspectral datasets create significant bottlenecks for clinical deployment. This chapter examines how edge computing and hardware acceleration technologies—including Field-Programmable Gate Arrays (FPGAs), Graphics Processing Units (GPUs), and specialized neural accelerators—enable real-time HSI processing at the point of care. We analyze architectural trade-offs, implementation strategies, and clinical workflow integration approaches that reduce processing times from minutes to milliseconds. Recent advances in embedded platforms achieve throughput rates exceeding 250 MSamples/s while maintaining power consumption below 15W, making intraoperative HSI guidance practically achievable.
Explainable Artificial Intelligence (XAI) constitutes a foundational pillar of responsible AI deployment, addressing the opacity of data-driven decision systems by making their internal logic transparent, auditable, and contestable. This chapter provides a systematic account of explainability and interpretability across the full spectrum of AI models — from classical supervised learning algorithms such as decision trees, support vector machines, and gradient-boosted ensembles to contemporary Large Language Models (LLMs) built on transformer architectures. Beginning with a structured taxonomy of XAI properties — scope, stage, model-dependency, and output type — the chapter critically examines canonical explanation methods including SHAP, LIME, attention visualization, mechanistic circuit analysis, and counterfactual generation.
Hallucination in large language models (LLMs) — the generation of content that is fluent and confident yet factually incorrect, logically inconsistent, or unsupported by available evidence — has emerged as one of the most consequential reliability challenges in contemporary AI development. As generative AI systems are deployed across high-stakes domains including clinical decision support, legal reasoning, scientific communication, and automated journalism, the consequences of undetected hallucination extend well beyond surface-level inaccuracy: they erode user trust, propagate misinformation, and undermine the foundational goal of building AI systems that are both truthful and aligned with human values. Addressing this challenge demands not only mitigation strategies but, crucially, robust and principled methods for detecting hallucination before, during, or after generation.