Scholarlink

<p>The combination of Generative Artificial Intelligence (GenAI) and Internet of Things (IoT) opens unexplored opportunities concerning the potential technological revolution and creates critical security concerns that requires systematic study. Based on refined 72 high-quality peer-reviewed studies published from 2017 to 2025, this systematic review adopts the rigorous Scientific Procedures and Rationales of Systematic Literature Reviews (SPAR-4-SLR) methodology to critically review these articles and provide two critical research questions: how GenAI can transform the capabilities, efficiency and robustness of IoT systems in diverse areas and what are the security concerns in the intersection of these two technologies. Through databases search in Scopus, Web of Science (WoS), and IEEE Xplore and further refinement of the search results with advanced bibliometric analysis software, including Visualisation of Similarities (VOS) viewer and Biblioshiny, this literature review shall be used to provide a compilation of the existing evidence-based research. The thematic analysis has found four thematic research clusters as Artificial Intelligence (AI) and Deep Learning (40% of publications), Cybersecurity and Adversarial Networks (28%), Intrusion Detection and Network Security (19%) and Anomaly Detection and Neural Networks (13%). The findings indicate that GenAI integration significantly changes the performance of the system, achieving 23–37% of predictive analytics improvement, 25% of energy savings and 29.2% reduction of latency. New security threats that convergence brings, including adversarial attacks, model poisoning, and privacy violations, are not covered by existing models of IoT security. These gaps that may be identified in the potential research are the lack of sufficiently scalable and energy-efficient AI models in resource-limited environments, inadequate structures of AI-specific threat mitigation and limited standardized evaluation approaches. The holistic roadmap for GenAI–IoT convergence provides a clear way forward for researchers, practitioners, and policymakers by identifying priority avenues for future research that are technologically driven, secure, and ethically conscious.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A systematic exploration of GenAI-driven security mechanisms for IoT in edge computing environments

  • 期刊论文

  • 发表日期：

    2026年5月2日

查看全文
  • Akansha Tripathi,
  • Jitendra Kumar Samriya

摘要

The combination of Generative Artificial Intelligence (GenAI) and Internet of Things (IoT) opens unexplored opportunities concerning the potential technological revolution and creates critical security concerns that requires systematic study. Based on refined 72 high-quality peer-reviewed studies published from 2017 to 2025, this systematic review adopts the rigorous Scientific Procedures and Rationales of Systematic Literature Reviews (SPAR-4-SLR) methodology to critically review these articles and provide two critical research questions: how GenAI can transform the capabilities, efficiency and robustness of IoT systems in diverse areas and what are the security concerns in the intersection of these two technologies. Through databases search in Scopus, Web of Science (WoS), and IEEE Xplore and further refinement of the search results with advanced bibliometric analysis software, including Visualisation of Similarities (VOS) viewer and Biblioshiny, this literature review shall be used to provide a compilation of the existing evidence-based research. The thematic analysis has found four thematic research clusters as Artificial Intelligence (AI) and Deep Learning (40% of publications), Cybersecurity and Adversarial Networks (28%), Intrusion Detection and Network Security (19%) and Anomaly Detection and Neural Networks (13%). The findings indicate that GenAI integration significantly changes the performance of the system, achieving 23–37% of predictive analytics improvement, 25% of energy savings and 29.2% reduction of latency. New security threats that convergence brings, including adversarial attacks, model poisoning, and privacy violations, are not covered by existing models of IoT security. These gaps that may be identified in the potential research are the lack of sufficiently scalable and energy-efficient AI models in resource-limited environments, inadequate structures of AI-specific threat mitigation and limited standardized evaluation approaches. The holistic roadmap for GenAI–IoT convergence provides a clear way forward for researchers, practitioners, and policymakers by identifying priority avenues for future research that are technologically driven, secure, and ethically conscious.