AI-Driven Automation Techniques for Enhanced Software Testing Efficiency

Abstract

Software systems especially large and complex ones has become increasingly difficult to manage with older, manual approaches. This paper investigates the use of artificial intelligence in improving software testing processes, not just for the sake of novelty, but because traditional methods are falling behind in real-world scenarios. Several AI-based tools and strategies are examined here, including approaches where machine learning generates test cases, algorithms that try to predict where defects will occur, and updated methods for regression testing that attempt to adapt over time. Rather than presenting AI as a silver bullet, the study compares these newer techniques with conventional testing methods. The evaluation considers how each performs in terms of detecting bugs, covering testable areas, and how much time or system resources are consumed during testing. One of the main contributions is a new model that combines reinforcement learning with natural language processing. This model isn't just theoretical it was applied to real-world projects to see how it would work in practice. The findings show that, in several cases, testing became faster, and more bugs were caught early, though results varied across different types of software. However, not everything worked perfectly. There were issues related to understanding how the AI made its decisions, and the quality of training data had a noticeable effect on outcomes. Also, fitting these tools into established software workflows wasn’t always smooth and required adjustment. Still, the overall takeaway is clear: AI has the potential to shift the field of software testing meaningfully, though there’s still work to be done to make these systems more interpretable and adaptable in real-time environments.

DOI: https://doi.org/10.17762/ijisae.v10i3s.7961