Katherine Elkins

Elkins attended Yale as an undergraduate, then completed a Ph.D. at UC Berkeley. She is the niece of Henry Elkins. == Teaching ==

Elkins is a professor of comparative literature and humanities in the Integrated Program for Humane Studies (IPHS) and faculty in Computing at Kenyon College. She is a founding co-director of the KDH lab and co-created the first human-centered artificial intelligence curriculum launched in 2016 at Kenyon College as the director of IPHS. Student and faculty research from the program has been downloaded over 90,000 times from institutions worldwide. Her recorded lectures with The Modern Scholar on The Modern Novel (2021) and The Giants of French Literature (2020) are tailored to broader public audiences via Amazon's Audible.com. She was nominated by graduating seniors to give the Kenyon College's December 2024 Baccalaureate Address which revisited David Foster Wallace's famous 2005 "This is Water" Kenyon graduation speech. In November 2025, Forbes featured the human-centered AI curriculum and research-focused AI Colab that Elkins co-founded with Jon Chun in 2016 within the Integrated Program for Humane Studies at Kenyon College. == Research ==

Elkins’s research addresses artificial intelligence in relation to literature, narrative, affective computing, and the ethics of AI. Her book The Shapes of Stories, published by Cambridge University Press in 2022, provided a comprehensive methodology for using diachronic sentiment analysis to analyze the emotional aspects of plot across dozens of literary classics using SentimentArcs. This method has been used to analyze narrative in diverse forms including literature, translations, TV scripts, end of life medical narratives, and the evolution of social media narratives for elections and economic crisis. Elkins was among the first scholars to analyze GPT-3's capabilities for creative writing. Her September 2020 paper 'Can GPT-3 Pass a Writer's Turing Test?' (co-authored with Jon Chun) -- published just four months after the model's release—was archived on gwern.net, a selective technical repository documenting AI development that is consulted by researchers at organizations including OpenAI and DeepMind. Her computational methodologies have been replicated in multiple studies archived alongside her work. Elkins' research addresses fundamental questions about information processing and transformation across different systems. Her work spans from analyzing how consciousness processes embodied information (Proust scholarship) to developing computational methods for preserving information in degraded cultural archives, identifying how AI systems process information differently than humans (syntactic framing vulnerabilities), and shaping governance frameworks for high-stakes decision-making. This includes analysis of how information transforms across languages, cultures, and media types, with consistent attention to what information is lost or preserved in each transformation. She presented the first transdisciplinary AI research at leading academic conferences including the Modernist Studies Association in October 2019, The International Society for the Study of Narrative in March 2020 and the Modern Language Association Conference in January 2021. Elkins was an early advocate for incorporating AI in literary studies with co-authored essays in The Journal of Cultural Analytics in September 2020 and Narrative in January 2021. More recently she focused on how AI redefines writing, creativity, authorship, translations of literature, eXplainable AI, and the future of the academia in leading journals like Poetics Today. Her article "A(I) University in Ruins: What Remains in a World with Large Language Models?" in the Proceedings of the Modern Language Association addresses how AI may fundamentally redefine traditional academic disciplines. Her collaborative position paper addressing the risks and benefits of open-source AI was selected for oral presentation at ICML in July 2024. Elkins traditional scholarship includes essays on Plato, Virginia Woolf, Franz Kafka, Marcel Proust, and William Wordsworth. In 2001 she won the A. Owen Aldridge Prize in Comparative Literature for an essay on Charles Baudelaire. Her 2025 human-centered AI scholarship includes "Beyond Plot: How Sentiment Analysis Reshapes Our Understanding of Narrative Structure" in the Journal of Cultural Analytics, "The Shapes of Cinderella: Emotional Architecture and the Language of Moral Difference" in the journal Humanities, and ""If Open Source is to Win, It Must Go Public" as a Spotlight talk at ICML 2025 CodeML Workshop. In December 2025, Elkins and her co-lead Jon Chun were awarded a grant of up to $330,000 from Schmidt Sciences as part of the Humanities and AI Virtual Institute (HAVI) for their project "Archival Intelligence: Rescuing New Orleans’ Endangered Cultural Legacy." One of only 23 such awards granted globally and totaling $11 million across all recipients, the 18-month initiative develops free, open-access AI tools to preserve and restore endangered materials in small community archives using smartphone photography, with a pilot focus on multilingual newspapers documenting Creole and Cajun communities, as well as early jazz artifacts in New Orleans, while addressing "cultural flattening" in AI models for multicultural and multimodal historical data. == Speaking ==

Elkins has given invited talks on AI, digital humanities, and higher education at institutions. As early as 2019, she publicly advocated integrating AI into traditional humanities curriculum with a keynote address at the Ohio State University. She gave the Meredith-Donovan lecture at Mount Saint Mary's University in 2023, featured AI Working Group lecture at Wofford College, and presentation at the Stories that Win Symposium at Washington University in 2024. Elkins gives keynotes on the intersection of AI, digital humanities, education, and the future of work. Most recently, in the summer-fall of 2024, these included keynotes at Carleton College's Day of Digital Humanities, Lafayette College AI Literacy Across the Curriculum, and Austin College's A.J. Carlson Lecture. Elkins speaks on a wide range of AI topics to diverse audiences both within the US and internationally. In February 2025 she was a featured speaker on "How agentic behavior, reasoning, and emotional intelligence upend previous notions of human exceptionalism" for the Khan Institute at Smith College which supports interdisciplinary research. That same month, she was the keynote for domestic and global public engaged action research hosted by Global Lab for the Global School at Worcester Polytechnic Institute. Elkins spoke on the connections between AI ethics, regulation, and interdisciplinary human-centered AI research at Rally Innovation 2025, a global cross-sector innovation conference that convenes visionary leaders, investors, entrepreneurs, educators, and other stakeholders. In October 2025, Elkins presented "The Power of Connection: Leveraging Technology for Humanistic Medical Education" for continuing medical education credits in Doha for Weill Cornell Medicine-Qatar. Also in October 2025, she was a featured speaker for the Education Guild in the OpenAI Forum in San Francisco, CA. Elkins has been a panelist on interdisciplinary AI conversations with thought leaders from diverse fields. She discussed language, epistemology and the ethics of AI with Ned Block, Francesca Rossi, and Dennis Yi Tenen in October 2022. Elkins debated AI generative art with co-panelist Boris Eldagsen (winner of Sony World Photography Award 2023) and Shane Balkowitsch on Al Jazeera in April 2023. She presented her perspectives on emotions at the intersection of AI and literature with experts Rosalind Picard, Joseph LeDoux, and Mabel Berezin. She discussed what gets lost in machine translation on the podcast Merging Minds. She is the AI industry expert for Bloomberg's new AI Strategy Course launched 2024. She serves as CAIO of HumanCentricLabs emphasizing humane applications of AI in the workplace. In December 2025, Elkins co-hosted an OpenAI forum on "AI and Academic Research" with Kevin Weil (VP of OpenAI for Science) and two other researchers in interdisciplinary human-centered AI including Leonardo Impett. == AI Safety and AI Regulation ==

In March 2024, Elkins was named a principal investigator in the U.S. AI Safety Institute Consortium, established by the National Institute of Standards and Technology (NIST) under Presidential Executive Order 14110. In June 2025, the consortium was renamed the U.S. Center for AI Standards and Innovation (CAISI), where she continues her safety research. She participates on behalf of the Modern Language Association (MLA), representing the organization and its 25,000 members as part of the inaugural safety research teams drawn from industry, academia, government, and non-profit sectors. Elkins co-authors research on AI governance and regulation. A position paper on international AI regulation, developed with international collaborators from institutions including the University of Oxford, ITS Rio (Brazil), UC Berkeley, Luxembourg IST, National University Philippines, Bocconi University, Tecnológico de Monterrey (Mexico), and others was selected for an oral presentation at the 2024 International Conference on Machine Learning (ICML). This work received support from META, and Elkins has been a member of META Open Innovation AI Research Community since 2023, and presented at the 2024 Conference at Meta's London Office. She also co-authored a comparative paper on AI regulatory frameworks in the United States, European Union, and China with the Oxford Witt Lab in 2025. With the organization PublicAI, she co-authored a piece arguing that open-source AI must be supplemented by public infrastructure to ensure models are accessible and governed in the public interest. Her work in applied AI ethics includes a project with the Notre Dame-IBM Tech Ethics Lab, which awarded her a grant for research into multi-agent court simulations for predicting juvenile recidivism. In January 2025, Elkins presented the findings from this grant-funded research at the University of Notre Dame. == References ==