Abstract

This talk explores how Large Language Models (LLMs) are transforming the software development lifecycle, from requirements engineering to design, implementation, testing, and beyond. While these models significantly enhance productivity and support developers in generating artifacts across all phases, they also introduce subtle and often overlooked risks, including hallucinated requirements, inconsistent designs, incorrect implementations, and misleading test cases. The seminar argues that LLMs should not be treated merely as coding assistants, but as active software components whose outputs must be carefully validated. Through real examples and demonstrations, the talk highlights how errors can propagate across the lifecycle and presents key challenges and open research problems in evaluating the correctness, consistency, and reliability of AI-generated software artifacts.

About the Speaker

Dr. Khaled Slhoub is an Associate Professor in the Department of Electrical Engineering and Computer Science at Florida Institute of Technology and serves as Program Chair for Computer Information Systems and Human-Centered Design. His research focuses on software engineering, particularly the use of AI-assisted tools, agent-based systems, and the evaluation and testing of AI and autonomous systems. He also works on frameworks for detecting disruptive behaviors in social bots and ensuring the reliability, fairness, and robustness of Generative AI models.