Computational Social Science + Humanities • AI Governance • Information Processing Research

Professor @KenyonCollege | Founder, KDH AI Lab
Principal Investigator @NIST US AI Safety Institute (representing Modern Language Association)| 90,000+ research downloads globally

I study information processing across political and cultural systems—investigating how information flows shape human behavior, social outcomes, and institutional decision-making. My work combines computational methods with interpretive depth, bridging social science and humanities to tackle questions about AI governance, behavioral prediction, and cultural influence.

Behavioral AI & Social Systems

• IBM-Notre Dame Tech Ethics Lab Grant ($60K): "How Well Can GenAI Predict Human Behavior?"
  • Collaborated with Yong Suk Lee (Notre Dame) on auditing LLMs for behavioral prediction accuracy
  • Multi-agent debate frameworks for improved explainability and stability in high-stakes decisions
  • Examining FATE (Fairness, Accuracy, Transparency, Explainability) in social applications

AI Governance & Policy

• Agentic: Computational approaches to AI alignment and policy simulation
• Hacking: Security vulnerabilities in large language models
• NIST AI Safety Institute: Federal AI safety frameworks (Principal Investigator, representing MLA)

Cultural & Political Analysis

• Archival Intelligence: AI methodologies for cultural heritage preservation
• GenAI for Narrative: Metrics and evaluation of narrative generation across multiple scales
• Hyperpersuasion: Evaluating hyperpersuasion in models
• Emotional Architecture: Information processing in narrative, media, and social systems

Computational Social Science Applications

• Supreme Court opinion analysis using sentiment and rhetorical patterns
• Union membership & wage correlation analysis by region
• Election polarization metrics (2022 midterms quantitative analysis)
• Social unrest visualization (Sri Lanka economic crisis case study)

Cultural & Behavioral Analytics

• Translation quality assessment using emotional equivalence metrics
• Human-chatbot interaction patterns and theory of mind in AI systems

Student Research Program

• 90,000+ downloads from 4,000+ institutions globally
• 300+ mentored projects spanning computational social science, digital humanities, and AI applications
• Topics: political discourse, economic modeling, cultural analytics, AI safety, behavioral prediction
• First human-centered AI curriculum globally (launched 2016)

Recognition

• Early GPT-3 researcher: 2020 analysis "Can GPT-3 Pass a Writer's Turing Test?" archived on gwern.net
• ICML contributor: Publications at intersection of AI and social science

Contributing to AI governance across multiple domains:

Federal Policy

• NIST US AI Safety Institute (Principal Investigator, representing Modern Language Association)
• Expertise: AI safety evaluation, human-centered frameworks, interdisciplinary risk assessment

International Frameworks

• UNESCO AI cultural preservation initiative
• Cross-national perspectives on cultural heritage preservation

Platform & Industry

• Meta Open Innovation AI Research Community
• OpenAI Higher Ed Forum Speaker
• Bloomberg AI Strategy Course (AI industry expert)
• RALLY tech innovation speaker

Open Source Advocacy

• International Public AI - Co-author: "If open source is to win, it must go public"
• Policy analysis of near to mid-term risks and opportunities in open-source generative AI

Computational Social Science & AI Policy

• "Near to Mid-term Risks and Opportunities of Open-Source Generative AI" (with Yong Suk Lee et al.) - ArXiv, 2024
• "If open source is to win, it must go public" - Policy paper on AI governance (International Public AI)
• "The Crisis of Artificial Intelligence: A New Digital Humanities Curriculum" (with Jon Chun) - Framework for humanistic social sciences + AI

Behavioral & Cultural Analysis

• The Shapes of Cinderella: Emotional Architecture and the Language of Moral Difference - Humanities (2025)
• "The Shapes of Stories: Sentiment Analysis for Narrative" - Cambridge University Press (2022)
  • First comprehensive methodology for diachronic sentiment analysis
  • Applications: literature, translations, TV, medical narratives, social media, political discourse

See full list: Google Scholar | Digital Kenyon

Why Computational Social Science + Humanities?

My training in comparative literature and humanities brings:

• Interpretive depth: Understanding how meaning-making shapes behavior (essential for studying misinformation, persuasion, political discourse)
• Cultural sophistication: Cross-cultural analysis skills that pure quant approaches miss
• Narrative expertise: How information structures influence decision-making in policy, economics, and social systems
• Multilingual capacity: Inform comparative policy and cultural analysis

Academic Training

• Ph.D., UC Berkeley
• B.A., Yale University
• Nearly 9 years hands-on AI implementation research
• Expertise: NLP, AI audits, social network modeling

Teaching Philosophy "Active engagement over passive critique" - Enable students to conduct cutting-edge computational research that tests theories across social sciences and humanities. Result: Globally downloaded research with real-world applications.

Academic Leadership

• Founded world's first human-centered AI Laboratory, KDH AI Lab (Kenyon College)
• Director, Integrated Program in Humane Studies (IPHS) - Kenyon's oldest interdisciplinary program
• Created first human-centered AI curriculum (2016)

• 🌐 Website: katherineelkins.com
• 💼 LinkedIn: Kate Elkins
• 📖 Google Scholar: Profile
• 🏛️ Institutional: Kenyon College
• 📚 Digital Kenyon: Research Repository
• 🐦 Twitter/X: @katelelkins