Integrating Machine Learning and Artificial Intelligence for Next-Generation Cybersecurity in Computer Science Applications

Abstract Cybersecurity remains one of the most pressing challenges in computer science as cyberattacks grow increasingly sophisticated, leveraging automation and adversarial intelligence. Traditional security mechanisms, including signature-based intrusion detection and rule-driven access control, often fail against zero-day exploits and advanced persistent threats due to their static nature and limited adaptability. Existing research in machine learning (ML) and artificial intelligence (AI) for cybersecurity has produced notable advancements, yet most frameworks remain constrained by isolated modeling, poor scalability, or high false-positive rates. To address these limitations, this study proposes a hybrid deep learning framework, CS-MLAI-Net, that integrates convolutional neural networks (CNNs) with bidirectional long short-term memory (BiLSTM) architectures, enabling both feature extraction and temporal attack pattern recognition. We evaluate CS-MLAI-Net on the NSL-KDD and CICIDS-2017 datasets, which encompass diverse modern intrusion types including denial-of-service, brute force, and infiltration attacks. Data preprocessing included normalization, categorical encoding, and synthetic minority oversampling to balance class distributions. Experimental results demonstrate superior detection accuracy of 98.7%, precision of 97.9%, and an F1-score of 98.2%, outperforming existing state-of-the-art methods by a significant margin while reducing false positives. The main contributions include a novel hybrid architecture tailored for cybersecurity, comprehensive benchmarking across multiple datasets, and robust preprocessing strategies to enhance generalizability. This work highlights the potential of AI-driven cybersecurity in securing next-generation digital infrastructures. Future research will extend CS-MLAI-Net towards real-time deployment in large-scale, cloud-native environments.