AI-Generated Test Automation for Autonomous Software Verification: Enhancing Quality Assurance Through AI-Driven Testing

Abstract

Test automation must be intelligent, scalable, and efficient due to the growing complexity of software systems. With the use of machine learning (ML), natural language processing (NLP), and reinforcement learning (RL), this study offers an AI-Generated Test Automation for Autonomous Software Verification that maximizes test case creation, defect detection, and execution speed. The suggested framework reduces execution time (110.7 ms) and resource use (310.5 MB) while improving test coverage (94.8%), defect detection rate (91.2%), and correctness (96.7%). The AI-driven method ensures minimal human interaction by automating the production of test cases, self-healing test scripts, and adapting to changing software modifications. The Full Model (Base + ML + NLP + RL) is the most effective method, with 98.2% test coverage, 99.4% accuracy, and 95.4 ms execution time, according to performance
comparisons of ML-based, NLP-based, RL-based, and combined AI-driven automation. According to the ablation study, hybrid AI models perform better than solo techniques in terms of fault discovery, testing effectiveness, and verification accuracy. Beyond software verification, a comparison of AI applications in radiology, heat pump optimisation, service sectors, and medicine demonstrates how AI affects a variety of fields. AI & AR in Radiology had the fastest processing speed (105.3 ms), AI in Medicine had the largest resource utilisation (360.7 MB), and AI in Service had the highest accuracy (94.1%). These results demonstrate
how AI may improve automation, decision-making, and performance optimisation in a variety of sectors. This study confirms that AI-powered test automation transforms quality assurance by lowering human testing efforts and improving software stability while guaranteeing scalability, dependability, and efficiency in Agile and DevOps contexts.