Historical knowledge introducing in the age of artificial intelligence: A study on deep research models
Abstract
This study examines the impact of artificial intelligence (AI) on historical research and the work of historians. It seeks to understand the transformations affecting the foundations and methodological principles of historical research in the age of digitization and AI. The latter provides rapid answers to various historical topics and issues, while AI tools have emerged that offer innovative solutions at every stage of historical research. Among these are intelligent models that utilize deep research capabilities, which can be employed to reimagine the past, construct historical facts, connect events, interpret incidents, and ultimately produce digital historical knowledge that rivals or surpasses human intellectual output.
The study's central problem revolves around the changes brought about by the AI revolution in the field of historical research and studies. It discusses the potential of deep research models to produce sound and rigorous historical knowledge and explores the relationship between human (natural) thinking and machine thinking in historical research. Is this relationship complementary or antagonistic, separate, and dysfunctional? Will the AI revolution overturn the rules and foundations of historical research methodology and ultimately render the traditional historian's profession obsolete?
This study's methodology is based on an examination of studies related to artificial intelligence (AI) in the humanities and social sciences. Its aim is to understand the relationship between historical knowledge, AI, and generative AI. We also explored the use of AI as an aid or alternative tool in historical research. An applied study was conducted on three generative AI models that operate using deep reasoning. The study was concluded by highlighting the various challenges of using AI in producing historical knowledge, particularly from ethical and methodological perspectives, as well as in terms of credibility, objectivity, and reliability. We also identified weaknesses and shortcomings in historical research produced using AI.
We targeted articles published in the past two years, as they reflect the developments in AI and its tools as aids in scientific and historical research in particular. We also relied on some classical studies on historical research methodology. We concluded with a number of findings, most notably that artificial intelligence offers promising opportunities in historical research, including gathering scholarly material, analyzing texts and sources, and translation. However, it faces limitations in accessing traditional (paper) sources in particular. Furthermore, it cannot replace the historian's role in producing sound historical knowledge characterized by depth of analysis, objectivity in presentation, and credibility and reliability in its findings.
Keywords: Digital historical knowledge; digital historian; generative artificial intelligence; deep research.
