Aug 2026· ACM Transactions on Software Engineering and Methodology· 0 citations· 68 references
TL;DR
GAI4-TDD makes failing tests pass with a success rate of about 90% on the first attempt; improves external/internal quality of software and students’ productivity; and professionals are generally-positive, although some barriers to GAI4-TDD adoption emerged.
Abstract
Test-Driven Development (TDD) is an agile practice that alternates writing failing (unit) tests (Red phase), implementing production code that makes failing tests pass (Green phase), and improving the code through the Refactor phase. TDD promises better external and internal quality of developed software and developers’ productivity, yet it remains difficult to practice. In prior work, some of the authors of this paper proposed integrating Large Language Models (LLMs) into the workflow of TDD, specifically during the Green phase by generating production code that makes failing (unit) tests pass. A supporting tool named GAI4-TDD (Generative Artificial Intelligence for Test-Driven Development) was also proposed to operationalize the integration of LLMs into the workflow of TDD for embedded systems development. In this paper, we validate this integration through a multi-study evaluation that is not limited to embedded systems development and comprises: (i) a laboratory study; (ii) a controlled experiment with graduate students in Computer Science/Engineering; and (iii) three industry-based qualitative studies. The results show that GAI4-TDD (i) makes failing tests pass with a success rate of about 90% on the first attempt; (ii) improves external/internal quality of software and students’ productivity; and (iii) professionals are generally-positive, although some barriers to GAI4-TDD adoption emerged.
Algorithm-Driven Development is introduced, a methodology developed from industrial practice to address recurring challenges in translating requirements into reliable, testable, and maintainable software behavior that provides systematic coverage of functional scenarios from the outset of development.
Philippe Jawish, Pierre Evrard, Alexandre Lemerle et al.· Journal of Systems and Softw...· 0 citations
An integrated reference architecture is proposed that combines TDD's fine-grained unit-level feedback loop with BDD's stakeholder-readable acceptance criteria inside a single continuous integration and continuous delivery (CI/CD) pipeline.
Urvish Gajjar· International Journal of Sci...· 0 citations
Context: Conversational Large Language Models (CLLMs) can automatically generate code by collaborating with users through natural language. However, poor collaboration can lead to poor quality output. Objective: This exploratory study aims to investigate how humans and CLLMs can collaborate as peers through vibe coding, an approach that integrates principles from prompt engineering, agile design, and human-AI co-creation to enhance collaboration. Method: We designed four interaction models representing different collaboration patterns in the software development process: the solo model (human-only development), the collaborative model (human-CLLM collaboration), the fully automated model (development autonomously performed by a CLLM), and the agentic model (development autonomously performed by the MetaGPT~X platform). Based on these models, we implemented corresponding Test-Driven Development (TDD) workflows using structured prompts and Python scripts. We then conducted a controlled pre-experimental study with TDD professionals to compare the solo and collaborative workflows. In addition, we performed repeated exploratory executions of fully automated and agentic workflows on the same development tasks to obtain complementary evidence. Results: Our findings suggest that the choice of interaction model should depend on the development objective. Agentic workflows are best suited for rapid development and functionally correct production code but may introduce additional implementation complexity. However, they may also introduce additional implementation decisions that are not explicitly required by the functional specifications, resulting in untested decision points. In contrast, collaborative workflows produce higher-quality, better-organized test suites. Conclusions: Our work explored how...
This paper presents a layered architecture for AI-integrated development systems, encompassing prompt formulation, semantic interpretation, code generation, and validation mechanisms, and explores the dynamics of human–AI interaction, highlighting the importance of trust calibration, control boundaries, and iterative refinement in achieving reliable outcomes.
Yasin Arik· International Journal of Res...· 0 citations
Modern software systems are characterized by continuous integration, frequent releases, heterogeneous architectures, and increasingly complex interaction patterns. These conditions place substantial pressure on conventional test-case design, particularly where manually authored tests struggle to achieve adequate coverage within constrained development cycles. This research examines an AI-enabled approach to test-case generation and optimization for modern software development by synthesizing evidence from studies concerning augmented reality, simulation-based learning, computational visualization, embedded-system monitoring, and AI-driven software quality engineering. The proposed methodology conceptualizes test generation as a pipeline comprising requirement interpretation, test-objective identification, candidate test generation, execution-oriented prioritization, redundancy reduction, and continuous optimization. Particular emphasis is placed on the relationship between intelligent automation and software quality engineering, where AI-driven frameworks can transform testing from a predominantly scripted activity into an adaptive quality-assurance process (Ramamurthy, 2023). The analysis indicates that AI can provide substantial benefits in generating diverse test scenarios, prioritizing high-value cases, and adapting test suites to changing software conditions. However, optimization effectiveness depends on the quality of requirements, training or heuristic signals, system observability, and validation mechanisms. The research therefore positions AI-enabled testing not as a replacement for engineering judgment but as an augmentation mechanism that improves scalability, coverage, and prioritization while retaining human oversight for critical decisions.
D. Perera, Nethmi Fernando· International Journal of Nex...· 0 citations
The integration of artificial intelligence (AI) tools into software engineering has fundamentally transformed development workflows, yet empirical evidence from developing economies remains scarce. This study investigates the impact of AI-assisted tools on software development lifecycle (SDLC) acceleration and code quality within Libyan enterprise environments. A comparative empirical study involved 48 professional developers across eight enterprises in Tripoli, Misurata, and Benghazi. Participants were stratified by experience level (junior, mid-level, senior) and assigned to either traditional manual development (TMD) or AI-assisted development (AIAD) using GitHub Copilot, ChatGPT, and Amazon Code Whisperer. Metrics included SDLC phase duration, code quality index, defect density, and perceived developer productivity. Results demonstrate that AI tools reduced development time by 22–38% across project scales, with junior developers realizing the greatest gains. However, AIAD projects exhibited a 12–21% increase in code churn, and security vulnerabilities were 1.21 to 2.10 times more frequent depending on project scale (1.21× for small projects, 1.80× for medium projects, and 2.10× for large projects). In large-scale projects, architectural inconsistencies attenuated benefits. Infrastructure constraints and digital literacy gaps moderated tool effectiveness in the Libyan context. These findings suggest that AI tool benefits are scale-dependent and require tailored governance frameworks for resource-constrained enterprises.
Suad Aghlilib· AlQalam journal of medical a...· 0 citations