METHODOLOGICAL ASPECTS OF IMPLEMENTING HYBRID GENERATIVE AI MODELS IN REGRESSION TESTING AUTOMATION PROCESSES

Abstract

The article analyzes the transformation of software quality
assurance practices under the growing adoption of Large Language Models (LLMs) in modern development
workflows. It examines the integration of generative AI into the lifecycle of automated regression testing,
including test design, page-object generation, implementation, and maintenance under UI changes. A
comparative experiment contrasts traditional manual automation with an AI-assisted approach using timebased and quality-oriented metrics such as time-to-feedback, maintainability indicators, and test stability.
The results demonstrate that a hybrid “human-in-the-loop” model significantly reduces effort in repetitive
automation tasks and accelerates initial test development, while also limiting technical debt growth through
faster adaptation to interface changes. At the same time, the study confirms the necessity of additional
verification stages - including prompt validation, static code analysis, and human review – to mitigate
hallucinated artifacts and manage the non-deterministic behavior inherent to probabilistic code generation.
Keywords: regression testing, Generative AI, LLM, test automation, Prompt Engineering, SDLC, QA
metrics, technical debt

References
1. Gregory J., Crispin L. Agile Testing: A Practical Guide for Testers and Agile Teams. Pearson
Education, Limited, 2021.
2. ISO/IEC/IEEE 29119-1:2022. Software and systems engineering — Software testing — Part
1: Concepts and definitions. URL: https://www.iso.org/standard/81291.html.
3. Garousi V., Felderer M., Mäntylä M. V. Guidelines for including grey literature and
conducting multivocal literature reviews in software engineering. Information and Software
Technology. 2019. Vol. 106. P. 101–121. URL: https://doi.org/10.1016/j.infsof.2018.09.006.
4. Capgemini, Sogeti. World Quality Report 2023–2024. URL:
https://www.capgemini.com/insights/research-library/world-quality-report-2023-24/.
5. Feldt R., Poulding S. Test Automation with Large Language Models. IEEE Software. 2023.
Vol. 40, no. 2. P. 18–26. URL: https://ieeexplore.ieee.org/document/10053992.
6. GitHub. Research: Quantifying GitHub Copilot’s Impact on Developer Productivity and
Happiness. 2023. URL: https://github.blog/research/publications/quantifying-github-copilotsimpact-on-developer-productivity-and-happiness/.
7. Selenium Documentation. The Page Object Model (POM) Design Pattern. URL:
https://www.selenium.dev/documentation/test_practices/encouraged/page_object_models/.
8. SonarSource. SonarQube Documentation: Maintainability and Technical Debt Metrics. URL:
https://docs.sonarsource.com/sonarqube/latest/user-guide/metric-definitions/.
9. Autify. Self-healing Test Automation: Concept and Workflow. URL:
https://autify.com/blog/self-healing-test-automation/.