Yasser Zarhloul Internship report

The development and use of personal assistants powered by Large Language Models (LLMs) are transforming technology, enhancing productivity for developers and general users alike. These AI tools efficiently assist in tasks like software development by automating code generation, which speeds up projects and reduces manual work. However, these capabilities carry risks related to intellectual property and licensing because LLMs train on large datasets, including code under restrictive licenses. Using such code could ex- pose companies to legal issues, including license violations with serious repercussions. To address these concerns, it is crucial for companies to use advanced scanning technologies to detect problematic code generated by LLMs. These tools, integrated within the software develop- ment lifecycle, must understand license nuances and identify similarities between generated and existing code to ensure compliance before deployment. This approach underlines the dual benefits and challenges of integrating sophisticated AI into everyday tools. It necessitates enhanced oversight and control to align with legal standards in soft- ware development, ensuring that companies can capitalize on the efficiency of AI while safeguarding against legal risks.