Katherine Elkins
Katherine Elkins is a Kenyon professor whose work connects AI research with literary interpretation and cultural archives. She earned a B.A. at Yale and a Ph.D. at UC Berkeley, and in 2016 she and Jon Chun founded Kenyon’s human-centered AI curriculum and AI CoLab. Her research tested GPT-3’s literary writing, compared models in The Shapes of Stories, and used AI to study translation and open-source governance. In 2025, she became project lead of Archival Intelligence; as of October 2026, she directs Kenyon’s Integrated Program in Humane Studies and co-leads the Modern Language Association team in NIST’s CAISI AI Consortium.
- Now at
- Kenyon College and NIST
Timeline
-
By 2016
Degrees and an AI curriculum
She holds a B.A. from Yale and a Ph.D. from UC Berkeley. After college, she lives in Paris, Prague and Asia before returning to the United States for graduate study. In 2016, she and Jon Chun found Kenyon’s human-centered AI curriculum and its AI CoLab.
-
2020
A writer’s Turing test
In 2020, she and Jon Chun publish “Can GPT-3 Pass a Writer’s Turing Test?”, which makes literary imitation and reasoning experimentally comparable.
-
2022
A book and full professorship
Her 2022 book The Shapes of Stories uses sentiment analysis, compares over three dozen models, and combines model selection with close readings of narrative passages. Kenyon’s trustees approve her promotion to full professor, effective July 1, 2022.
-
2023
An AI curriculum for humanities
In 2023, she and Jon Chun publish “The Crisis of Artificial Intelligence: A New Digital Humanities Curriculum for Human-Centred AI” in the International Journal of Humanities and Arts Computing. The article outlines a human-centered AI digital humanities curriculum and ways it opens research paths for the humanities and humanistic social sciences.
-
2024
Translation and open-source AI
In her 2024 study “In Search of a Translator: Using AI to Evaluate What’s Lost in Translation,” she applies stylometry, emotional arcs and multilingual models such as GPT-4o and Mistral to compare translations of Proust’s In Search of Lost Time. She also coauthors “Position: Near to Mid-term Risks and Opportunities of Open-Source Generative AI,” an ICML paper that maps risks and benefits of open and closed generative AI and argues for responsible open release.
-
2025
Archival Intelligence in New Orleans
Schmidt Sciences selects her to lead Archival Intelligence in 2025, a project building an open AI pipeline for endangered New Orleans archives. It uses fragile newspapers, sheet music, letters and recordings to study the emergence of jazz.
-
2026
An ICML spotlight on public AI
In 2026, she coauthors “If Open Source Is to Win, It Must Go Public,” a position paper accepted as an ICML Spotlight. The paper argues that open-source AI needs public infrastructure and institutions to make models accessible, sustainable and governed in the public interest.
-
As of October 2026
Kenyon professor and NIST co-lead
As of October 2026, she is a professor of comparative literature and humanities and director of Kenyon’s Integrated Program in Humane Studies. She co-leads the Modern Language Association team in the NIST CAISI AI Consortium and serves as a co-principal investigator of Archival Intelligence with Jon Chun.
Career
- Since 2016Kenyon CollegeCo-founder, Full Professor
- NowNISTCo-lead of Modern Language Association team
Education
Sources and credits
Photo: We have no published photo of Katherine Elkins yet. Send us a free-licensed photo.
- Katherine Elkins | Kenyon College
- On the record: Katherine Elkins on Artificial Intelligence
- Katherine Elkins — AI researcher
- The Shapes of Stories
- Full Professors | Kenyon College
- The Crisis of Artificial Intelligence: A New Digital Humanities Curriculum for Human-Centred AI
- In search of a translator: using AI to evaluate what’s lost in translation
- Position: Near to Mid-term Risks and Opportunities of Open-Source Generative AI
- Schmidt Sciences awards $11M in grants to bring AI to humanities research
- Archival Intelligence — Rescuing New Orleans’ Endangered Cultural Legacy
- If open source is to win, it must go public