Somayeh Fatahi

Project Summary: Nowadays, with the advancement of technology, especially artificial intelligence, we are witnessing the emergence of novel tools. In the latter part of 2022, OpenAI introduced ChatGPT, an AI tool built on top of existing Large Language Models (LLMs), to facilitate interactive communication through a conversational interface. Notably, ChatGPT exhibits the capacity to generate diverse forms of text, encompassing scientific abstracts, domain-specific question responses, programming code, lifelike conversational exchanges, summarizations, translations, suggestions, and recommendations. In addition to all the capabilities of ChatGPT, it also has drawbacks like the possibility of spreading fake information. We aim to compare the responses of ChatGPT to those generated by humans. The study underscores potential risks associated with AI-generated content being mistaken for human-written text, raising concerns about misinformation. The research objective involves a comparative analysis to evaluate the quality, accuracy, and stylistic differences of ChatGPT's responses compared to human-generated answers. To evaluate our idea, we utilize a dataset containing programming code written by programmers, as well as the responses generated by ChatGPT. This project opens the door for studies addressing  ChatGPT's strengths and limitations, user preferences, and the challenges of differentiating between AI-generated and human-written content.

Internal Mentor: Dr. Chanchal Roy and Dr. Julita Vassileva

External Mentors: Dr. Aaron Genest, Siemens Company