Evaluation of Artificial Intelligence Models in the Field of Islamic Philosophy: ChatGPT-Google Bard

Abstract

Islamic Philosophy, which originated in the Hijaz Peninsula, developed as a discipline through interactions with various civilizations. During the early periods of Islam, especially under the Umayyads and Abbasids, influences from Greek, Indian, Persian, and Chinese cultures were incorporated, and these influences were synthesized into Islamic philosophy in a unique manner. Islamic Philosophy has been deeply intertwined with various scientific fields such as logic, metaphysics, medicine, and astronomy, and over time, it has had a significant impact on Western philosophy. In the 21st century, artificial intelligence (AI) models have been rapidly utilized in fields like science and social sciences, playing a key role in knowledge acquisition via internet-based platforms. ChatGPT and Google Bard, with their human-like language processing capabilities, have generated significant interest, and this study aims to assess the performance of these two models in the field of Islamic Philosophy. Methodology: In this study, answers to 5 open-ended questions on Islamic Philosophy were obtained from ChatGPT and Google Bard. These questions encompassed a variety of content areas, such as information retrieval, comparison, and analysis. The responses were evaluated based on criteria like clarity, specificity, relevance, accuracy, supporting information, consistency, sufficiency, integrity, and data volume. This evaluation was presented in a comparative table, and the results were discussed. Findings of the Study: Some of the responses to the questions on Islamic Philosophy and their evaluations are as follows: What is Islamic Philosophy? ChatGPT provided a general definition of Islamic philosophy, referencing interactions with Greek, Indian, and other cultures, but did not sufficiently emphasize the unique aspects of Islamic philosophy. Google Bard offered a more comprehensive explanation, connecting Islamic philosophy to issues in theology, reason-revelation, metaphysics, ethics, and politics. Bard also included a wider range of references, while ChatGPT relied more on Western sources. In the question of the influence of the Neoplatonic understanding of the universe on Suhrawardi's theory of light, both AI models address similar topics. ChatGPT generally examines the hierarchy of being, the relationship between light and knowledge, and the matter-soul relationship, while Google Bard adopts a more systematic and detailed approach. Bard delves deeper into topics such as cos mology, unity, the light metaphor, and the nature of knowledge. However, both models fail to adequately address Suhrawardi's unique contributions to the theory of knowledge. As a result, the provided information is superficial, and there are gaps in the analytical analysis required for these topics. Provide Information on the Introduction of Islamic Philosophy Courses into the Turkish Education System: ChatGPT provided a detailed account of the evolution of Islamic philosophy courses in the education system from the Ottoman period to the Republic era. It elaborated on the contributions of philosophers like Farabi, Avicenna, and Al-Ghazali in Ottoman philosophical studies. Google Bard, however, provided a more concise explanation, stating that philosophy in the Ottoman period was limited. ChatGPT also explained in more detail how philosophical courses were integrated into new educational institutions following the closure of madrasas in the Republican period. Compare the Ideal State Theories of Plato and Farabi: Google Bard adopted a more systematic and analytical approach in comparing the ideal state theories of Plato and Farabi, outlining their similarities and differences. Bard highlighted the shared belief in the ruler being a philosopher-king, the importance of education, and the necessity of social justice. It also noted the differences in their views on social hierarchy-Plato's rigid class system versus Farabi's more flexible approach-and briefly touched on how their views on religion and cosmology influenced their political theories. ChatGPT, on the other hand, provided a more surface-level comparison, with a few sentences, and made an incorrect claim about Farabi's focus on individual happiness in his political theory. In general, both AI models provided useful information on Islamic Philosophy, but they lacked depth in certain areas. Google Bard presented a broader perspective, offering a more comprehensive analysis of various philosophical systems and cultural contexts. ChatGPT, on the other hand, was limited to shorter, more straightforward answers, but sometimes provided more detailed explanations on specific topics. In a complex and extensive field like Islamic Philosophy, both models sometimes offered incomplete or misleading information. Thus, users should critically evaluate both sources. The study compared the performance of ChatGPT and Google Bard in answering questions on Islamic Philosophy. The results showed that both AI models provided accurate but incomplete and occasionally incorrect answers. Google Bard demonstrated a stronger performance with more analytical, detailed explanations, references, and a deeper exploration of the topics. ChatGPT, however, was more limited, offering shorter responses and less comprehensive sources. Overall, while both models provided helpful insights, their information was insufficient for advanced study. The study concludes that strategies need to be developed to enhance the effectiveness of AI models in specialized disciplines like Islamic Philosophy.