The emergence of Large Language Models (LLMs) has profoundly reshaped computational linguistics, enabling unprecedented reasoning, context awareness, and semantic understanding capabilities. Integrating these sophisticated models into Internet-of-Things (IoT) ecosystems holds transformative potential for enabling intelligent, autonomous, and contextually-aware applications. This article begins with an extensive state-of-the-art survey of existing literature on the integration of LLMs within IoT environments, establishing foundational insights into current capabilities, limitations, and deployment frameworks. Subsequently, the manuscript contributes a comprehensive analysis of lightweight LLMs and embedding models suitable for resource-constrained IoT platforms while introducing a taxonomy of sub-billion–parameter ( < 1B), mid-range (1B–2B), and exact 2B–parameter LLMs—spanning families such as Qwen, Llama, SmolLM, and IBM’s Granite—as well as embedding models under 1B parameters optimized for low-latency retrieval. Comparative assessments elucidate trade-offs in model size, inference latency, context windows, energy consumption, and performance across models categorized by parameter count. Next, a diverse spectrum of prospective use cases—including home healthcare, smart agriculture, industrial optimization, and environmental monitoring—demonstrates the practical efficacy of deploying tailored LLM-IoT frameworks for real-world problem-solving. Later, the article systematically explores key challenges that must be addressed to fully realize the integration of LLMs within IoT contexts, encompassing resource constraints, heterogeneous data processing, privacy and security risks, latency requirements, model interpretability, and ethical considerations. Finally, we outline critical directions for future research, advocating advancements in IoT-specific model architectures, multimodal sensor fusion strategies, real-time adaptive inference methods, energy-aware inference scheduling, and privacy-preserving federated learning paradigms.
Federated Learning (FL) is a machine learning paradigm that enables collaborative model training across multiple devices, such as those found in the Internet of Things (IoT), while preserving data privacy. Despite its potential, FL is vulnerable to attacks, including data poisoning. This paper introduces an Adaptive Privacy-Preserving FL (APPFL) method that helps mitigate these risks. APPFL adjusts the influence of clients by adaptively weighting each update, ensuring that the contribution of each client is dynamically adjusted to improve accuracy. It incorporates local differential privacy to enhance individual data privacy further. The efficacy of APPFL is evaluated using simulated IoT devices and various datasets, including MNIST and CIFAR-10, demonstrating its robustness against poisoning attacks and its ability to maintain privacy. This research contributes to the ongoing efforts to secure FL, a critical technology in today’s data-driven industries.
Internet of Everything (IoE) is an extension of Internet of Things (IoT) that has revolutionized the digitalization of our society by connecting things and humans to create a digital world. As a result, IoE has reshaped the way that humans and autonomous systems interact and exchange information in network connections. However, the digitization imposes new security, trust, and safety concerns, such as the risk of undesirable behavior of driverless cars, which might cause dangerous road consequences. In this context, Deep-Learning (DL) applications have emerged in trust management to proactively improve the efficiency, reliability, security, trust, and safety of different IoE entities, and then achieve the sustainability of different social IoE relationships. To support further progress in integrating DL-based solutions in trust management, this paper reviews the existing literature related to DL-based trust management. Specifically, the review targets the DL-based approaches employed in trust management, their components and application areas in the context of IoE, and trust relationships in IoE. Furthermore, it highlights a set of research challenges, opportunities and recommendations related to the adoption of proactive trust management using deep learning in IoE.
Large Language Models (LLMs) are rapidly transitioning from standalone conversational systems to autonomous agents that reason, plan, and interact with external tools. While this shift enables powerful applications in domains such as healthcare, finance, law, and software engineering, it also introduces new security and safety risks. Attacks such as prompt injection, jailbreak exploits, backdoor triggers, and multimodal adversarial inputs expose vulnerabilities not only at the model level but also across the broader agentic workflow. Existing defenses—ranging from input filtering and alignment reinforcement to runtime monitoring—remain fragmented and often fail to anticipate adaptive adversaries. Meanwhile, red teaming has emerged as a critical methodology for stress-testing these systems; however, current efforts lack standardization, comprehensive coverage across modalities, and integration with agent-specific contexts. This paper provides the first comprehensive survey of LLM agent security, synthesizing research on attack strategies, red teaming frameworks, evaluation suites, and defense mechanisms. We categorize automated and agentic red teaming approaches, highlight domain-specific vulnerabilities in code, web, and multimodal agents, and analyze defense strategies spanning prompt-level, decoding-time, runtime, backdoor, privacy-preserving, and multi-agent safeguards. Building on this synthesis, we outline key open challenges and future research directions, including the need for scalable defenses, standardized benchmarks, robustness against adaptive attacks, explainability, and secure integration of multi-agent workflows. Our findings aim to guide both researchers and practitioners in advancing robust, trustworthy, and resilient LLM-powered agents for safety-critical applications.
The smart Oil and Gas Industry (OGI) is a transformation of the corresponding automated industry via using Industry 4.0 (I4.0) enablers, including Internet of Things (IoTs), cloud computing, blockchain, big data analytics, robotics, simulation, 3D printing, augmented reality/ virtual reality, system integration, and cybersecurity. IoTs facilitate and accelerate real-time data gathering using proper sensors in any conditions and then transmit the collected data set to cloud servers via the internet. This paper contributes to assessing the OGI adaptation to IoT-based management by proposing an assessment framework involving assessment criteria and the corresponding hierarchical structure. In addition, the proposed framework includes the Analytical Hierarchy Process (AHP) to derive the significance of criteria and sub-criteria based on experts’ judgment. Results of employing this framework in a case study from Iran showed that the application of IoT in some sectors, such as pipeline and Health/Safety/Environment (HSE), is crucial, and in other sectors of OGI, like Exploration and production, is less significant, according to the questioned experts. In addition, emergency maintenance, monitoring pipeline parameters, real-time detection of dangerous conditions, and well automation are the most significant sub-criteria of maintenance, pipeline, HSE, and exploration and production, respectively.
