HUMAN Fellows

Anya Golovkova

Vibe Philology: Translation of the Glory of the Goddess

This project develops an innovative approach to translation at the intersection of philology, ethnography, and artificial intelligence. Borrowing from the idea of vibe coding, ‘vibe philology’ names a mode of translation that integrates close reading with affective, contextual, and AI-assisted interpretation. The project centers on producing a new translation of the Devi Mahatmyam (Glory of the Goddess), a foundational Sanskrit text (c. sixth–eighth centuries CE) that remains widely recited in contemporary ritual contexts across South Asia. While the text has been translated multiple times, existing translations largely privilege textual and theological analysis, often neglecting its lived, performed, and affective dimensions. This project asks: how might translation better capture not only semantic meaning, but also ritual “vibe”—the sonic, embodied, and interpretive textures through which texts are experienced in practice?

Paul Henne

How Ordinary People Think About Artificial Intelligence 

As AI becomes more common in everyday activities, we need to understand how people interpret AI systems in relation to human beings and how those judgments shape ethical questions about responsibility, trust, and the role of AI in human life. Using the methods of experimental and analytic philosophy, I examine how people make causal judgments about AI and humans in familiar, everyday scenarios rather than only in unusual hypothetical cases like the moral dilemmas that often appear in philosophical research. This comparison will allow me to test whether people judge human and artificial agents differently when thinking about causation

Justin Kee

Four Souls in the Machine

This project aims to experiment with and analyze AI cross communication utilizing work by Mary Shelley, Phillip K. Dick, William Gibson, and Ludwig Wittgenstein. Student researchers develop philosophical charters that defines AI agents’ values, epistemic style, and cultural orientations, and serve as real-time curators of the agents’ interactions, deciding when to inject prompts, provocations, or redirections, and when to let a line of machine reasoning play out to its conclusion. Researchers then perform close readings of the complete debate transcripts, identify patterns of interpretive divergence across agents.

Don Meyer

American Dark Ages: Further Explorations into Generative AI Applications for a Multi-Media Production

Last year’s Human Fellows grant generously supported a multi-media project called American Dark Ages, set in a post-apocalyptic America some 450 years from now. This project consists of three interlocking sets of three—a trilogy of traditional fantasy novels, each of which will also appear as a graphic audiobook and as an album of original music heard in the audiobook. Last year, my students and I focused on the application Eleven Labs for voice-acting and sound effects, which strives to support artists who participate in their “voice cloning” applications. As of late April 2026, we have nearly completed both the sound effects and character voices for the first novel. This year, in addition to continuing this voice-over work with Eleven Labs, we will be exploring the potential for AI to help edit the text and to generate audio and video material for a trailer for the 
project.

Mel Rovner

Critical Analysis of Architectural History Pedagogy

This project applies computational and humanistic methods to examine how architectural history has been constructed, canonized, and taught—and to intervene in that process through critical pedagogy. Drawing on my ongoing research into AI visual analysis and NLP-based text analysis of archival and institutional documents, I analyze a corpus of architecture history syllabi drawn from the Society of Architectural Historians' teaching resources. Using large language models and computational image analysis across multiple AI APIs, I will conduct close reading at scale, examining which buildings are represented, how they are described, which geographies and cultures are centered, and what aesthetic and ideological assumptions structure the canon. This constitutes both a literature review of the field, asking what architecture historians have collectively decided is worth teaching, and a form of textual and visual interpretation that uses AI as an instrument of critical analysis rather than neutral extraction.

Dan Schwartz

Generating the “Global Sound”: Cultural Appropriation, Stereotype, and AI Music Creation

This project investigates how generative AI constructs and commodifies “global” musical styles. Situated within the World Music Survey course, the project examines whether AI-generated music reproduces long-standing patterns of exoticism, simplification, and cultural appropriation rooted in colonial histories and global market forces. Students use text-to-music AI platforms to generate compositions based on prompts referencing specific musical cultures and then compare with historically grounded recordings and ethnomusicological scholarship. Students analyze how AI systems flatten musical complexity, exaggerate sonic signifiers, or detach music from ritual, language, and community context.