The Role of Artificial Intelligence and Machine Learning in DevSecOps: Leveraging Predictive Analytics, Automated Threat Detection, and Anomaly Identification for Secure Software Delivery Pipelines

Abstract

This study investigates the integration of  artificial intelligence (AI) and machine learning (ML) within DevSecOps frameworks to enhance secure software delivery pipelines through predictive analytics, automated threat detection, and anomaly identification. A mixed-methods research design was employed, combining quantitative analysis of a simulated dataset comprising 150,000 CI/CD pipeline logs (2018–2022) with qualitative insights from 12 expert interviews. Key findings reveal that ML-driven anomaly detection reduced false positives by 68% and improved threat prediction accuracy to 92.4%. Predictive models identified 87% of vulnerabilities prior to deployment, while automated remediation decreased mean time to patch (MTTP) from 14.2 hours to 2.1 hours. The study concludes that AI/ML integration significantly strengthens DevSecOps maturity, enabling proactive security without compromising velocity. These outcomes underscore the transformative potential of intelligent automation in achieving security-as-code at scale.