IEEE Internet of Things Magazine • June 2022 46 GUEST EDITORIAL W ith the advent of the Internet of Things (IoT), vari- ous industries have made considerable progress [1], including agriculture, utilities, manufacturing, and retail. IoT solutions help to increase productivity and efciency in factories and workplaces. Meanwhile, in smart cities, intercon- nected trafc lights and parking lots are established through IoT technologies to alleviate traffic problems. It has been possible thanks to recent advances in IoT frameworks and protocols [2]. Currently, IoT security has become the main concern in practical applications [3]. In the field of Artificial Intelligence of Things (AIoT), artifcial intelligence (AI) is realized through various tech- nical supports, especially machine learning (ML) technology. ML can analyze data, learn from them, and infer or predict with specific algorithms. ML can solve practical problems that are difcult for simple rules using data and is widely used in complex tasks, such as search engine (SE), autonomous driving (AD), machine translation (MT), medical diagnosis, spam fltering, com- puter games, face recognition, data matching, credit rating, and even IoT trafc classifcation [4]. In particular, ML technology can shape the surrounding environment into a personal preference under IoT, which is very practical. IoT security threats, such as distributed denial of service (DDoS) attack, extortion software, and social engineering, can steal the key data of individuals and organizations. Hackers can attack IoT infrastructures over the Internet through secu- rity vulnerabilities. Moreover, there is a need for IoT intrusion detection systems based on ML [5]. Many communication tech- nologies can be used over IoT, such as WiFi, Bluetooth, and IoT cards. IoT cards are sure to spread throughout IoT as a basic communication means among IoT devices. Convenience services have come along with information security threats with the appearance of IoT. Eavesdropping can hardly be avoided in traditional cable and optical fber communication even through software encryption. In the feld of 5G + AIoT, the number of intelligent terminals is enormous, most of which have security vulnerabilities, such as weak passwords, mutual attacks, and neighboring infections. These IoT security problems should be solved from multiple perspectives and comprehensively [6]. Specifcally, safety regulations, design, standards, technologies, and management should be improved and deepened. Vari- ous methods can be adopted in AIoT to reject miscellaneous attacks. The mechanisms of most attacks are to send data ran- domly over IoT and get unanticipated results. Forms of attacks are diverse, including collision attacks, blocking attacks, and information tampering, hindering the development of IoT. Secu- rity authentication identifes intelligent devices based on secu- rity identity to prevent unauthorized access and forgery, thus improving security from the source. Operators provide the main authentication with the 5G network unifed authentication framework to ensure legitimate access. Slicing authentication is carried out through slice ID to avoid unauthorized user occu- pation of slicing resources. The combination of encryption, authentication technology, and multi-channel transmission has a very broad application prospect for the secure communication of multi-channel transmission. It is a new solution to security problems such as anti-intermediary attacks and key cracking. With the popularization of 5G, IoT technology will be extended to various industries vertically, and a new marketing space will be created. Meanwhile, the importance of IoT securi- ty issues cannot be overemphasized. The frst article by Abdur Rahman et al. addresses the concept of the cyber entity in the era of 6G. 6G will allow billions of AIT devices and networks to connect at a speed of less than 0.1 ms. It is expected to provide a speed of terabits per second. With all these advancements, the future generation of smart applica- tions will relate to people, processes, and critical infrastructures. This will also open the door to vulnerabilities of the connected systems. Hence, ensuring the security of the cyber entities is of utmost importance. Deep learning (DL) systems will be integrat- ed with the 6G systems in which the network and the nodes that will form the 6G will have native intelligence to handle the secu- rity threats. They will be expanding their research with ensembles of DL methods, explaining the ability to handle further levels of security threats. The second article by Singh et al. addresses patient digital records, known as electronic health data (EHD), which have security and privacy concerns in the existing IoT-based health- care systems. They explore the promising but very challenging research area of IoT-enabled watermarking schemes for health- care applications. Further, a comprehensive review of current studies related to IoT-based healthcare with watermarking tech- niques, along with their merits and limitations, is provided in the article. The contributions of reviewed approaches are then summarized and compared from diferent technical perspectives. Lastly, they detail and comment on recent challenges, along with several directions of potential research. SECURE COMMUNICATIONS OVER THE INTERNET OF ARTIFICIALLY INTELLIGENT THINGS: PART 2 Zhihan Lv Jaime Lloret Houbing Song Jun Shen Wojciech Mazurczyk