Year: 2025
Type: preprint
Abstract: Mutation testing is a fault-based software testing technique that evaluates test suite effectiveness by introducing small changes, called mutants, into program code and measuring whether existing tests can detect them. Despite its strong theoretical foundations, traditional mutation testing faces adoption barriers, including high computational cost, equivalent and redundant mutants, and the manual effort required to write new tests. Recent advances in Artificial Intelligence (AI), particularly in machine learning, deep learning, and large language models (LLMs), offer new opportunities to address these challenges. AI can enhance mutation testing through intelligent mutant generation, predictive mutant selection, equivalent mutant detection, and automated test case generation. This paper provides a detailed survey of AI-driven techniques applied to mutation testing, reviews prominent academic and industrial case studies such as DeepMutation, MOTIF, and Meta’s Automated Compliance Hardening (ACH) system, and outlines future research directions. By integrating AI into the mutation testing workflow, the process can become more selective, context-driven, and automated, enabling its practical use in large-scale, real-world software engineering projects.
Authors: Pankaj Bhandula, Nityanand Rai
Institution: Amazon (United States)
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4413638604