Google or ChatGPT: Who is the Better Helper for University Students

Overview

• This study investigates the academic help-seeking preferences of Taiwanese university students, comparing their use of ChatGPT and traditional search engines like Google.
• The researchers used a mixed-methods approach to understand the key factors influencing students' preference for using ChatGPT for academic assistance.
• The study found that students generally tend to prefer using ChatGPT over Google when seeking academic help, highlighting the potential popularity of generative AI models in the educational field.
• The researchers also compared the performance of various machine learning algorithms and identified Random Forest and LightGBM as the most effective in predicting the factors influencing students' help-seeking preferences.

Plain English Explanation

This study looked at how college students in Taiwan choose to get academic help online. The researchers wanted to see if students are more likely to use the artificial intelligence chatbot ChatGPT or the traditional search engine Google when they need help with their studies.

The researchers used a mixed-methods approach, which means they combined different research methods like surveys and data analysis, to understand the key reasons why students prefer to use ChatGPT for academic assistance.

The study found that overall, students tend to choose ChatGPT over Google when they need help with their schoolwork. This suggests that generative AI models like ChatGPT are becoming increasingly popular in the education field.

The researchers also compared different machine learning algorithms to see which ones were best at predicting the factors that influence students' preferences for using ChatGPT. They found that the Random Forest and LightGBM algorithms performed the best.

Technical Explanation

The study employed a mixed-methods research design, combining quantitative and qualitative approaches, to investigate Taiwanese university students' online academic help-seeking preferences. The researchers surveyed 312 students to collect data on their usage of ChatGPT and Google for academic assistance.

To analyze the survey data, the researchers compared the performance of seven machine learning algorithms, including Logistic Regression, Decision Tree, Random Forest, XGBoost, LightGBM, Multilayer Perceptron, and Support Vector Machine. The Random Forest and LightGBM algorithms exhibited the best predictive performance and were subsequently used to evaluate the influence of 18 potential factors on students' help-seeking preferences.

The analysis revealed that three core factors significantly influenced how university students sought academic help online: ChatGPT fluency, ChatGPT distortions, and age. These findings suggest that as students become more fluent in using ChatGPT and aware of its limitations, they are more likely to rely on the AI chatbot for academic assistance rather than traditional search engines.

Critical Analysis

The study provides valuable insights into the shifting preferences of college students when it comes to seeking academic help online. However, there are a few limitations to consider:

1. The research was conducted in Taiwan, so the findings may not be fully generalizable to other cultural contexts. Generative AI models may be perceived and utilized differently in different educational systems and regions.

2. The study focused on university students, leaving out the perspectives of younger or older learners. The factors influencing help-seeking preferences may vary across different age groups.

3. The researchers did not explore the quality or effectiveness of the academic assistance provided by ChatGPT and Google. Further research is needed to assess the actual impact of these tools on student learning outcomes.

Despite these limitations, the study highlights the growing importance of understanding how generative AI models are transforming academic help-seeking behavior. Educators and policymakers should consider these findings when developing strategies to foster critical thinking and responsible use of technology in the classroom.

Conclusion

This study provides valuable insights into the academic help-seeking preferences of Taiwanese university students, revealing a growing preference for using the AI chatbot ChatGPT over traditional search engines like Google. The researchers identified key factors, such as ChatGPT fluency and awareness of its limitations, that influence this trend.

These findings have important implications for the education sector, as they underscore the need for educators to prioritize the development of critical thinking skills in students, while technical personnel should work to enhance the fluency and reliability of ChatGPT and other generative AI models. Integrating chat and search functions may also help achieve an optimal balance in academic help-seeking.

Overall, this study highlights the rapid evolution of information technology tools for academic assistance and the importance of understanding how students are adapting to these changes.