Evaluating AI-Assisted Search vs. Traditional Keyword Search: A Comparative Study of ChatGPT and Google
Owen F. Karpinski, Levi N. Mcfetridge, Michael C. Ta, Faris N. Ashriem, Aditi Singh · 2025
This paper presents a comparative analysis of AI-assisted search using ChatGPT and traditional keyword-based search via Google. Unlike conventional search engines that require users to browse through links, ChatGPT delivers direct, conversational responses. Both methods are evaluated across 80 diverse, manually curated queries, focusing on response time and factual accuracy (hallucinations). Results indicate that ChatGPT offers faster answers—approximately 2.5 times quicker on average, but is more prone to factual inaccuracies. In contrast, Google provides more reliable and verifiable information, albeit with longer response times. Key contributions of this study include: a systematic evaluation of ChatGPT and Google across multiple query types, a quantitative comparison of speed and factual correctness, and an exploration of the trade-offs between user efficiency and information reliability. Findings suggest that while AI-assisted search enhances user convenience, traditional search engines remain superior in terms of factual trustworthiness. A hybrid search model combining the speed of AI with the accuracy of traditional search engines is proposed to optimize future information retrieval.