INTEGRATION OF ARTIFICIAL INTELLIGENCE INTO THE SOFTWARE DEVELOPMENT LIFE CYCLE (SDLC)
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
This article examines the transformation of the Software Development Life Cycle (SDLC) under the influence of generative artificial intelligence tools integration. The objective of this study is to conduct a comparative analysis of the efficiency and security of AI assistants (specifically GitHub Copilot, Google Gemini, GPT-4) across all development stages: from requirements gathering to deployment and maintenance (AIOps). The research methodology includes systematic literature analysis, classification of contemporary AI tools by type (generative, predictive, analytical ML), and conducting a practical experiment comparing GPT-4 and Google Gemini 2.5 Pro models based on code correctness, security (SQL Injection vulnerabilities), and propensity for package hallucinations. The evolution of development methodologies from the classical Waterfall model through agile approaches (Agile, DevOps) to the modern AIOps paradigm, where artificial intelligence performs autonomous monitoring, failure prediction, and system self-healing, has been analyzed. The study reveals that AI integration fundamentally transforms the developer's role from syntax writing to architectural oversight, generated code verification, and AI agent orchestration. The ability of modern models to generate secure code using parameterized queries has been experimentally confirmed; however, critical risks of non-existent library hallucinations (Package Hallucinations) have been identified, creating a Supply Chain attack vector through the Slopsquatting mechanism. Particular attention is devoted to intellectual property issues concerning AI-generated code, risks of confidential data leakage through Shadow AI, and the Vibe Coding phenomenon leading to degradation of fundamental skills among novice developers. The transition to the Agentic AI concept, where software development transforms into a process of managing autonomous specialized agents, has been substantiated. The necessity of implementing new security protocols, generated content verification, and revision of educational programs for training a new type of specialists – AI orchestrators – has been emphasized.
Fields of Study & Subjects
Software Engineering Research
Software Engineering Techniques and Practices
Artificial Intelligence Applications
Publication details
- Journal / source
- Cybersecurity Education Science Technique
- Volume
- 4
- Issue
- 32
- Pages
- 376–376
- Publication date
- 26 March 2026
- Language
- EN