Year: 2025
Type: conference-paper
Abstract: Artificial Intelligence (AI) has shown promise in accomplishing tasks traditionally performed by humans across a range of domains. However, while previous studies have assessed AI's ability to perform these tasks, there remains a gap in understanding whether AI can carry out these tasks with the same level of expertise and finesse as human professionals. Moreover, while AI's performance has been examined in certain domains, it has not been sufficiently tested across all fields where human expertise is vital. To address these gaps, we conducted a study involving 32 expert participants across various domains, who provided prompts and evaluated the responses generated by ChatGPT Plus. Our results, consistent across domains, reveal that AI can perform tasks at a basic level, most of the time, but it falls short in replicating the depth and quality of human expertise. Despite these limitations, AI can provide a solid starting point and assist experts in specific contexts. Future work can explore the use of other forms of AI, such as humanoid AI bots, to assess tasks that require interaction among multiple participants or teams, offering potential avenues for deeper collaboration and more advanced automation.
Author: Jad A. Atweh
Institution: University of Virginia
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4411086913