LLMs as Mirror, Colleague, Rival

CALL FOR ABSTRACTS

Large language models (LLMs) have quickly become a prominent feature of contemporary intellectual and cultural life, raising distinctive questions for scholars across the digital humanities and related disciplines. We are interested in the multifaceted role of LLMs in academic research. LLMs process and generate language in a way that is both familiar and uncanny, revealing and opaque. They can write, translate, argue, and create, but also lead us astray. In their complexity, they hold up a strange mirror to human thought and culture (to borrow Shannon Vallor’s metaphor).

This symposium takes as its organizing metaphor three roles that LLMs play in (digital) humanities research: as mirror, colleague, and rival. As a mirror, LLMs reflect the values and biases encoded in training data drawn from a large corpus of human-generated text. Studying the output of LLMs (and how it falls short) can teach us about ourselves as wellasthe technology itself. As a colleague, LLMs can serve as research tools or co-authors, raising questions about collaboration, authorship, research integrity, and the evolving nature of scholarly work. As a rival, LLMs can disrupt and confound, challenging the epistemic foundations of academic research, by undermining replicability and evaluation, and flattening the research landscape.

These three roles are not mutually exclusive, and the tensions between them are precisely what makes LLMs a productive object of study for digital humanists, philosophers, communication scholars, cultural theorists, cognitive scientists, and others working adjacent to the digital humanities alike.

Guiding Questions

This symposium aims to deepen our understanding of the role of LLMs in (digital) humanities research, focusing on questions such as:

• What can LLMs teach us about human language, cultural heritage, knowledge, and creativity?
• In what ways do LLMs encode or distort cultural values, biases, and worldviews? How can our disciplines help us identify and critique these?
• How can scholars productively collaborate with LLMs as research tools? What methodological and ethical issues does this raise?
• What does the rise of LLMs mean for domain expertise and the division of cognitive labor in the (digital) humanities?
• What normative and political questions are raised by the delegation of linguistic and cognitive tasks to LLMs?
• How do LLMs function as rivals or obstacles in (digital) humanities research?  In what ways can they undermine traditional research methods and standards?
• How do the geopolitics of LLM development and deployment affect their use in academic research (e.g., in terms of academic freedom, conflicts of interest)?

We aim to answer these questions from a variety of disciplinary perspectives. We welcome theoretical, empirical, and methodological contributions. We invite speakers to present on a broad range of topics including, but not limited to the cognitive and AI (e.g., modelling of individual and collective cognition, LLMs as human subjects, the nature of LLMs more broadly construed), arts and media (e.g., shifting definitions of authorship; the potential dispossession of artists from creative industries), philosophical (e.g., LLMs and value-sensitive design, cognitive deskilling, chatbot epistemology and ethics), linguistic (e.g., modeling language acquisition and processing, corpus annotation and analysis), and communication and information studies (e.g., the role and risks of chatbots in domains of health, information, and well-being; the contribution of LLMs to social and digital inequalities; the integration of LLMs into communication science methodologies).

Submitted abstracts ideally (but not necessarily) feature digital humanities methods or reflect on digital media and technologies.

This 2-day, hybrid symposium – part on-site in Tilburg, part online – brings together scholars from a range of disciplines (all represented in the Tilburg School of Humanities and Digital Sciences) to engage in a cross-disciplinary dialogue on these matters.

Submission Guidelines

We invite interested speakers to submit (i) an anonymized abstract of max. 300 words, and (ii) a cover sheet including your name,  institutional affiliation, and whether you would prefer to give a talk in person or online to DHsymposium@tilburguniversity.edu by May 1st, 2026. You’ll be notified on May the 22^nd.