Technologies such as artificial intelligence (AI), blockchain, and the Internet of Things (IoT) have converged in driving the next wave of digital revolution. Amalgamating the aforementioned advancements with critical infrastructure (CI) can significantly help society by offering a quality of life and boosting the nation’s economy and productivity. However, the lack of cybersecurity in CI gave rise to advanced threats and vulnerabilities that hindered the aforementioned societal benefits. In this vein, the paper provides an in-depth analysis of cyber threats and risks associated with different critical infrastructures, such as the financial, agriculture, energy, and healthcare sectors. Further, we thoroughly investigate the staggering benefits of AI and, based on it, present an exhaustive solution taxonomy to showcase the competency of AI mechanisms in confronting cyberattacks on CI. The taxonomy specifically addresses issues like data privacy, algorithmic bias, and human-AI collaboration for CI. Further, we proposed an AI-based secure data exchange framework for smart grid CI, where we attempt to secure the sensor’s data (i.e., power consumption, energy readings, and network data) from malicious adversaries. The proposed framework is evaluated using statistical measures, such as accuracy, training time, and receiver operating characteristic (ROC) curve, and anomaly detection. Further, the paper examines the research challenges that still adhere to the critical systems and require stringent AI-based mechanisms to tackle them.

人工智能(AI)、区块链和物联网 (IoT) 等技术已融合在一起，推动下一波数字革命。将上述进步与关键基础设施 (CI) 相结合可以通过提供生活质量并促进国家经济和生产力来显着帮助社会。然而，CI 中网络安全的缺乏导致了高级威胁和漏洞，阻碍了上述社会效益。本着这一精神，本文深入分析了与金融、农业、能源和医疗保健等不同关键基础设施相关的网络威胁和风险。此外，我们深入研究了人工智能的惊人优势，并在此基础上提出了详尽的解决方案分类法，以展示人工智能机制在应对 CI 网络攻击方面的能力。该分类法专门解决了 CI 的数据隐私、算法偏差和人类与人工智能协作等问题。此外，我们提出了一种基于人工智能的智能电网 CI 安全数据交换框架，我们试图保护传感器数据（即功耗、能源读数和网络数据）免受恶意对手的侵害。使用统计指标来评估所提出的框架，例如准确性、训练时间、受试者工作特征（ROC）曲线以及异常检测。此外，本文还探讨了仍然坚持关键系统并需要严格的基于人工智能的机制来应对的研究挑战。