AI TOOLS IN SCIENTIFIC DISCOVERY – TRANSFORMING RESEARCH PARADIGMS

Artificial intelligence (AI) is increasingly transforming the landscape of scientific research by providing advanced tools and approaches that enhance hypothesis generation, data analysis, and experimental modeling. Despite ongoing concerns regarding AI’s broader societal impacts, its application in science offers tangible benefits, enabling researchers to identify patterns in large and complex datasets that are difficult for humans to discern, formulate new hypotheses based on existing studies, and conduct simulations that accelerate the testing and validation of ideas. This study presents a systematic review of scientific literature and AI tool documentation used in research, drawing on data from Scopus, Web of Science, and Semantic Scholar, as well as practical examples of AI system applications across various scientific domains. The analyzed tools include large language models (LLMs) such as ChatGPT, GPT-4, and Claude; Elicit; Scite.ai; ResearchRabbit; and machine learning platforms including Scikit-learn, TensorFlow, and PyTorch. Research methods involve a systematic literature review to identify trends, a comparative analysis of tools in terms of functionality and applicability, and an examination of practical cases of AI use for hypothesis generation, data analysis, and modeling. Results indicate that AI substantially facilitates five key areas: (1) idea generation and hypothesis formulation through literature and citation analysis; (2) detection of patterns and correlations in large datasets; (3) automation of literature review and visualization of scientific networks; (4) AI-driven simulations and generative modeling for predicting outcomes and exploring new experiments; and (5) logical consistency checking of hypotheses against existing data and theoretical frameworks. While AI does not fundamentally redefine the scientific method, it significantly increases research efficiency, broadens the scope of scientific inquiry, and opens new avenues for discoveries previously inaccessible. The rapid evolution of AI tools creates a dynamic environment for their integration into scientific practice, necessitating continuous assessment of their capabilities, limitations, and ethical implications.