Abstract: Connecting objects with sensors and actuators over wired or wireless networks is the goal of the Internet of Things (IoT). The Internet of Things (IoT) is predicted to link more than 25 billion devices by 2020. Data received from these devices will rise tremendously in the years to come. Because of the variability in response time and geographic location, IoT devices generate a large amount of data in a variety of formats. AI calculations may play an important role in ensuring security and approval when biotechnology and IoT frameworks are combined, an unusual finding when working on security and usability. Assailants, on the other hand, frequently use learning calculations to exploit the flaws in cutting-edge IoT frameworks. In this research, we use AI to identify spam in IoT devices, resulting in a better level of security. Machine learning-based IoT spam detection is given in order to accomplish this goal. Using a range of metrics and data sets, five AI models are analysed in this system. On the basis of the revised input highlights, an overall spam score is produced for each model. The reliability of an IoT device is evaluated by this score. A new strategic strategy is given the go-ahead thanks to the REFIT Smart Home dataset. The findings show that the proposed conspiracy is a viable alternative to the existing schemes.

| DOI: 10.17148/IARJSET.2022.96112