CHRP Digital Humanities Workshop Series 2 - Beyond Tools and Critics: Human-Machine Synergy in Literary AI
Arts & Culture | Seminar | CCL24250582
Abstract:
When a sentiment analysis algorithm labels a metaphor of loss as ‘neutral,’ or a chatbot confidently misattributes a myth, we are witnessing more than technical error—we are confronting a fundamental divergence in how humans and machines interpret culture. This talk moves beyond debates about artificial intelligence (AI) as a tool or adversary, instead proposing a framework for collaboration where humanistic reasoning and machine logic productively collide. Through projects spanning Irish conflict poetry, comparison of traditional algorithms and latest large language models (LLMs) for sentiment analysis, and re-reading poetry by archival retrieval, this seminar explores how gaps between human and machine interpretation can spark innovation in both scholarship and industry.
By dissecting case studies like how different computational methods misclassify the sentiments of poetry of violence and trauma and a retrieval-augmented generation (RAG) system that surfaces historical context for silenced political verses, this sharing demonstrates how humanities scholars and AI do not just coexist—they co-create. Machines expose patterns invisible to close reading; humans inject cultural nuance and historical gravity into algorithmic outputs. Together, they challenge assumptions about efficiency, objectivity, and the ‘correct’ way to read.
Beyond the academy, this synergy holds transformative potential for industries built on narrative. Examples include collaborations with filmmakers employing AI not to automate storytelling, but to surface overlooked connections in plot structure, and how Irish mythological figures inform the making of AAA games. These partnerships underscore a radical proposition: Humanities training—with its emphasis on ambiguity, context, and critique—is a strategic asset in building AI systems that resonate with human audiences.
The talk concludes by urging a reevaluation of the humanities’ role in an AI-driven world. If machines excel at scale and pattern recognition, humanists excel at asking what patterns matter and for whom. By framing our disciplines not as critics of technology, but as architects of human-machine collaboration, we can shape AI that amplifies—rather than flattens—cultural complexity.
When a sentiment analysis algorithm labels a metaphor of loss as ‘neutral,’ or a chatbot confidently misattributes a myth, we are witnessing more than technical error—we are confronting a fundamental divergence in how humans and machines interpret culture. This talk moves beyond debates about artificial intelligence (AI) as a tool or adversary, instead proposing a framework for collaboration where humanistic reasoning and machine logic productively collide. Through projects spanning Irish conflict poetry, comparison of traditional algorithms and latest large language models (LLMs) for sentiment analysis, and re-reading poetry by archival retrieval, this seminar explores how gaps between human and machine interpretation can spark innovation in both scholarship and industry.
By dissecting case studies like how different computational methods misclassify the sentiments of poetry of violence and trauma and a retrieval-augmented generation (RAG) system that surfaces historical context for silenced political verses, this sharing demonstrates how humanities scholars and AI do not just coexist—they co-create. Machines expose patterns invisible to close reading; humans inject cultural nuance and historical gravity into algorithmic outputs. Together, they challenge assumptions about efficiency, objectivity, and the ‘correct’ way to read.
Beyond the academy, this synergy holds transformative potential for industries built on narrative. Examples include collaborations with filmmakers employing AI not to automate storytelling, but to surface overlooked connections in plot structure, and how Irish mythological figures inform the making of AAA games. These partnerships underscore a radical proposition: Humanities training—with its emphasis on ambiguity, context, and critique—is a strategic asset in building AI systems that resonate with human audiences.
The talk concludes by urging a reevaluation of the humanities’ role in an AI-driven world. If machines excel at scale and pattern recognition, humanists excel at asking what patterns matter and for whom. By framing our disciplines not as critics of technology, but as architects of human-machine collaboration, we can shape AI that amplifies—rather than flattens—cultural complexity.
Registration
Registration Through SLES