Loading article...
Loading article...
Stream/Area:
-
Subject:
-
Paper ID:
IJSREM68679
Volume:
10 |
11 |
Nov 2026
Pages:
1-7
Publication Frequency:
Monthly (12 issues per year)
License:International Journal of Scientific Research in Engineering and Management follows the Creative Commons Attribution 4.0 (CC BY 4.0) License, allowing free use and distribution with proper attribution.
Department of Master of Computer Application
JCEI's Jaihind Institute of Management and Research, Kuran, Pune
(Assistant Professor),Department of Master of Computer Application
JCEI's Jaihind Institute of Management and Research, Kuran, Pune
The rapid proliferation of Large Language Models (LLMs) specialized in programming languages has fundamentally reconfigured the modern Software Development Life Cycle (SDLC). Far beyond basic autocompletion, generative foundation models now actively assist software practitioners across requirements formulation, architectural drafting, automated unit test synthesis, vulnerability remediation, and continuous refactoring. This empirical investigation provides an in-depth, rigorous evaluation of generative AI's real-world impact across engineering teams. Analyzing telemetry across 50 production engineering cohorts and standard code evaluation suites (HumanEval, MBPP, and SWE-bench), we demonstrate that modern generative coding assistants accelerate routine boilerplate authoring by 52.1% to 57.8%, yet introduce pronounced cognitive friction during pull request verification. Crucially, synthetic benchmark pass rates (such as HumanEval Pass@1 topping 92.4%) mask an acute architectural reasoning deficit when resolving multi-file repository-level bugs on SWE-bench Lite (38.8% resolve ceiling). Furthermore, automated static security audits reveal that unconstrained generative synthesis increases latent defect density by 14.6%, primarily manifesting as improper input validation (CWE-20), SQL injection vectors (CWE-89), and hallucinated third-party dependencies. To resolve this dilemma, we formulate an end-to-end verified SDLC architecture integrating Abstract Syntax Tree (AST) validation, Retrieval-Augmented Generation (RAG) repository indexing, and mutation adequacy gating, achieving high delivery velocity without sacrificing code hygiene.
KEYWORDS
Generative AI represents a transformative technological inflection point in software engineering, offering unprecedented productivity acceleration across routine coding and testing tasks. However, realizing its full potential requires addressing the acute gap between syntactic fluency and architectural robustness, alongside proactive governance against code security vulnerabilities and reviewer fatigue. Future research frontiers must center on self-healing software architectures, multi-agent cooperative verification swarms, and federated codebase learning models that preserve corporate proprietary confidentiality while continually refining domain-specific code synthesis.
[1] S. Peng, E. Kalliamvakou, P. Cihon, and M. Demirer, "The Impact of AI on Developer Productivity: Evidence from GitHub Copilot," arXiv preprint arXiv:2302.06590, 2023. [2] A. Ziegler, et al., "Measuring GitHub Copilot's impact on productivity," Communications of the ACM, vol. 67, no. 3, pp. 54-63, 2024. [3] X. Hou, et al., "Large language models for software engineering: A systematic literature review," ACM Transactions on Software Engineering and Methodology, vol. 33, no. 5, pp. 1-47, 2024. [4] DeepSeek-AI, "DeepSeek-Coder: When the Large Language Model Meets Programming," arXiv preprint arXiv:2401.14196, 2024. [5] C. E. Jimenez, et al., "SWE-bench: Can Language Models Resolve Real-World GitHub Issues?," in Proc. ICLR, 2024, pp. 1-28. [6] M. Chen, et al., "Evaluating Large Language Models Trained on Code," OpenAI Tech. Rep., arXiv:2107.03374, 2021. [7] N. Perry, et al., "Do Users Write More Insecure Code with AI Assistants?," in Proc. ACM CCS, 2023, pp. 2785-2799.
Publication Date
11/2/2026
Peer Review Type
Double-Blind Peer Review
Review Rounds: 2
Documents
Article Views
0
Total views since publication
PDF Downloads
3
Total downloads since publication
Geographic Reach
5
Digital Preservation: This article is permanently archived on Zenodo and complies with all data preservation standards across major indexing platforms.
Plagiarism Screening
This manuscript has been checked for plagiarism using TurnitinPlagiarismChecker software.
Conflict of Interest
The authors declare no conflict of interest related to this research work.
Funding Information
No funding information available for this research.
Related research in the same subject area
Loading similar articles...