In the quest to unravel the complexities of human morality, recent research has taken an innovative turn by leveraging artificial intelligence to analyze ancient taboo lists, offering a fresh perspective on the foundational categories that shape moral judgment. This groundbreaking study, conducted by Y. Feder and published in Humanities and Social Sciences Communications, revisits one of the oldest sources of moral guidance—the prohibitions etched into the cultural fabric of ancient civilizations—and attempts to distill them into coherent moral categories through a data-driven, linguistic approach. The implications challenge prevailing psychological frameworks while highlighting AI’s emerging role in deepening our understanding of moral cognition.
The research interrogates the conventional efforts to compartmentalize morality into distinct, evolutionarily grounded categories, such as those proposed by Moral Foundations Theory (MFT) and the Moralized Authority Construct (MAC). These frameworks have long sought to identify innate psychological mechanisms underlying moral judgments by interpreting behavioral repertoires shaped over millennia. However, Feder’s study pivots away from purely top-down theoretical models and instead applies AI tools to analyze the empirical content of taboo lists from various early human societies—items that have withstood the test of time as markers of what was considered off-limits in different cultural milieus.
Central to the investigation is the development of what is termed the Taboo-Based Inventory (TBI), a newly proposed framework distilled from logical and linguistic analyses of ancient taboos. Rather than fitting data into pre-existing theoretical molds, the TBI framework emerges directly from the patterns and commonalities found within the taboos themselves. This method represents a bottom-up approach, where categories are deduced from the lived realities and prohibitions recorded by ancient peoples, rather than inferred evolutionary intentions. The result is a set of moral categories that provide more nuanced alignment with the taboo items under scrutiny compared to categories suggested by MFT or MAC.
To validate these findings, the study employed a multidisciplinary experimental task wherein both human participants and AI systems independently categorized taboo items according to different frameworks. Remarkably, the TBI categories yielded a significantly higher consensus in classification, implying a closer fit to how the content of taboos is mentally organized and perceived. This suggests that the moral distinctions ancient societies encoded were coherent and potentially universal in ways that pre-arranged theoretical categories might not fully capture.
Importantly, the study refrains from declaring the TBI superior or more valid than existing moral frameworks. Instead, it positions its categories as provisional and subject to continuous empirical scrutiny and validation. The notion of “validity” in this context is tied to the frameworks’ ability to detect evolutionary embedded moral intuitions—a criterion the TBI does not explicitly seek to satisfy. Feder stresses that the TBI’s purpose is descriptive rather than evolutionary explanatory, making it a complementary lens rather than a competing theory.
One of the study’s most provocative contributions is the challenge it presents to traditional evolutionary moral psychology. By demonstrating that ancient taboo behaviors—arguably some of humanity’s earliest moral expressions—do not neatly conform to MFT or MAC categories, the research calls for a reassessment of how these frameworks incorporate historical and ethnographic data. Ancient taboos embody moral concerns that have evolved over tens of thousands of years and ignoring them risks overlooking critical facets of the moral psyche’s formation. Evolutionary models, Feder argues, must reconcile with this evidence to produce a more comprehensive account of morality’s origins.
The significance of ancient taboo lists extends beyond mere historical curiosity. Their content reflects fundamental concerns about social cohesion, purity, and harm that persist in contemporary moral discourse. Despite arising in religious or ritualistic contexts, these taboos convey principles that remain interpretable—and relevant—across cultures today. The study highlights how these enduring prohibitions can be viewed as a form of proto-moral inventory, providing a window into universal moral building blocks that underpin human societies.
The deployment of artificial intelligence in this cultural and moral archaeology exemplifies a transformative shift in humanities research. AI’s pattern-recognizing capabilities enable the extraction of latent structures from large, complex textual datasets that would be difficult to parse through human analysis alone. In Feder’s study, AI tools facilitated the identification of coherent categories that would otherwise remain obscured within the vast and nuanced corpus of taboo items. This interaction between human cognitive insight and machine-assisted analysis represents a powerful new frontier in decoding the intricacies of human moral motivation.
Moreover, this research underscores AI’s potential beyond technical or scientific domains, showcasing its ability to illuminate fundamental aspects of human self-understanding. As moral decisions guide every facet of social life, the ability to dissect their underpinnings could inform disciplines ranging from ethics and anthropology to psychology and artificial moral agents’ design. These insights could prompt novel approaches to moral education, conflict resolution, and even the development of AI systems capable of navigating complex ethical landscapes.
The study also invites ongoing dialogue between descriptive and normative approaches to morality. While evolutionary theories often aim to explain why certain morals persist or how they arose, the TBI’s descriptive categories derived from tangible behavioral taboos provide a grounded mapping of what is actually prioritized across societies. Future research could explore how these descriptive moral contours interact with normative ideals and whether integrating the two perspectives could yield richer, more holistic theories.
In sum, Feder’s investigation presents an innovative fusion of ancient cultural wisdom, artificial intelligence, and moral psychology, probing the foundational elements that govern human behavior from a new vantage point. By focusing on the taboo lists of antiquity, this research not only enriches our understanding of morality’s historical depth but also pushes the boundaries of current theoretical frameworks, encouraging scholars to account for evidence long recorded in human prohibitions.
As the ethical landscapes of modern societies evolve—often accelerated by technological advancements and globalization—returning to the primal roots of moral reasoning offers a sobering reminder of the shared human concerns that transcend time and geography. The study’s interdisciplinary approach exemplifies how integrating AI, linguistic analysis, and cultural history can yield novel insights, positioning future inquiry at the crossroads of tradition and innovation.
The dialog between AI models and human subjects in this research also raises fascinating questions about the nature of moral cognition itself. If both can reliably categorize intricate taboo items according to the same set of derived principles, it suggests a compatibility in the representational structures of morality that transcends biological implementation. This opens the door for pioneering collaborations between cognitive science and artificial intelligence to further map the moral mind.
Finally, while Feder emphasizes caution in interpreting these categories’ ultimate validity, the research undeniably moves the conversation forward by introducing a robust, data-driven framework that draws directly from humanity’s ancestral moral landscapes. It serves as a compelling invitation to reconsider what moral psychology might learn from ancient prohibitions and what AI might yet reveal about the ethical foundations shared by all human beings.
Subject of Research: Using AI to analyze ancient taboo lists to identify coherent moral categories and assess their implications for existing moral psychology frameworks.
Article Title: Using AI to identify moral categories in ancient taboo lists
Article References:
Feder, Y. Using AI to identify moral categories in ancient taboo lists.
Humanit Soc Sci Commun 12, 1171 (2025). https://doi.org/10.1057/s41599-025-05509-0
Image Credits: AI Generated