<p>Intrusion detection systems (IDS) play a crucial role in safeguarding Internet of Things (IoT) systems by detecting and identifying malicious or unauthorized activities. The rapid proliferation of IoT devices has increased the requirement for robust network security. IoT and advanced IoT technologies are essential for 6G networks, giving dense connectivity, low latency, ultra-reliability, and higher performance. Artificial intelligence (AI), including deep learning (DL) and machine learning (ML), presents solutions for enhancing and leveraging innovative technologies in next-generation radio communications. Moreover, security remains a demanding problem that is resolved using ML and DL-based intrusion detection models. In a world where 6G IoT infrastructure growth is a global priority, ensuring robust intrusion detection is vital to defend sensitive data and ensure the seamless process of critical models. In this paper, the Artificial Intelligence-Based Intrusion Detection and Secure Communication (AIBID-SCSA) technique is proposed. The main goal of the AIBID-SCSA technique is to enhance the accuracy and efficacy of attack recognition in an assisted 6G-IoT network. Initially, the AIBID-SCSA technique applies min-max normalization to measure the input data into a uniform format. Furthermore, improved sparrow search algorithm (ISSA)-based feature selection is employed to identify the most relevant feature from the network traffic data. For the classification of intrusion detection, the long short-term memory (MIX_LSTM) model is used. Finally, the rabbit optimization algorithm (ROA) model is implemented for the hyperparameter selection process to optimize the detection results of the MIX_LSTM model. The experimental validation of the AIBID-SCSA method is investigated under the TON_IoT_Train_Test_Network dataset. The comparison analysis of the AIBID-SCSA method portrayed a superior accuracy value of 99.63% over recent state-of-art techniques.</p>