| Title: | The hidden cost of artificial intelligence in programming |
| Authors: | Shibaev, I. S. |
| Keywords: | материалы конференций;large language models;programming;synthetic data;model collapse;feedback loops;hallucinations |
| Issue Date: | 2026 |
| Publisher: | БГУИР |
| Citation: | Shibaev, I. S. The hidden cost of artificial intelligence in programming / I. S. Shibaev // Электронные системы и технологии : сборник материалов 62-й научной конференции аспирантов, магистрантов и студентов БГУИР, Минск, 13–17 апреля 2026 г. / Белорусский государственный университет информатики и радиоэлектроники ; редкол.: П. В. Камлач [и др.]. – Минск, 2026. – С. 1131–1132. |
| Abstract: | This article examines a structural contradiction in AI-assisted programming. While large language models accelerate code generation, their improvement is constrained by the exhaustion of high-quality human-written training data. As synthetic data increasingly enters training pipelines, recursive learning from model-generated content may create feedback loops leading to model collapse, reduced code diversity, and hallucinations in software development. |
| URI: | https://libeldoc.bsuir.by/handle/123456789/64616 |
| Appears in Collections: | Электронные системы и технологии : материалы 62-й конференции аспирантов, магистрантов и студентов (2026)
|
