2026-05-25T14:37:43Zhttps://keep.lib.asu.edu/oai/request

oai:keep.lib.asu.edu:node-2001072025-04-01T19:36:44Zoai_pmh:alloai_pmh:repo_items

200107 https://hdl.handle.net/2286/R.2.N.200107 http://rightsstatements.org/vocab/InC/1.0/ http://creativecommons.org/licenses/by-nc-sa/4.0 2025-05 29 pages Smaw, Jacob Roumina, Kavous Diaz Lopez, Andres Barrett, The Honors College Department of Information Systems Large language models have rapidly transformed productivity and information retrieval since the release of ChatGPT in November 2022. These AI models excel at understanding and generating human language, often mimicking logical thinking with vast knowledge access. While their capabilities are impressive, ranging from writing and debugging code to providing career advice, they remain limited in areas such as emotional understanding, common sense, and originality. This thesis examines the strengths and weaknesses of LLMs through nine structured tests conducted on ChatGPT and Co-Pilot, as these are among the most widely used models. Three experts, Professor Alan Simon from Arizona State University, along with Bill Jirik and Raphael Leviton from Eide Bailly LLC, provided analysis based on their expertise in data science and AI. The study presents tests in a progressive order based on performance rankings, offering insights into AI’s real-world applications and limitations. Understanding these strengths and weaknesses is crucial as AI continues to integrate into daily life, helping businesses, developers, and users make informed decisions on its practical use. Artificial Intelligence Large Language Models Living up to the Hype? A Challenging Exercise for ChatGPT and Co-Pilot