The use of Large Language Models (LLMs) in OSS security poses challenges, including the potential for LLMs to misinterpret security reports, generate incomplete or inaccurate patches leading to regressions, and the uninterpretability of LLM decision-making processes which causes hesitance to trust AI-generated suggestions.

Authors

• Person: Not available Organization: arXiv
  A Mixed-Methods Study of Open-Source Software Maintainers On ...

Sources

• A Mixed-Methods Study of Open-Source Software Maintainers On ... arxiv.org arXiv via serper

Referenced by nodes (1)

• Large Language Models concept