Maarten Sap - Home (original) (raw)
I am an assistant professor at CMU's LTI department with a courtesy appointment in HCII, and a part-time research scientist and AI safety lead at the Allen Institute for AI (AI2). My research focuses on (1) measuring and improving AI systems' social and interactional intelligence, (2) assessing and combatting social inequality, safety risks, and socio-cultural biases in human- or AI-generated language, and (3) building narrative language technologies for prosocial outcomes.
I received my PhD from the University of Washington where I was advised by Noah Smith and Yejin Choi.
[bio for talks]
My research group:
Overarching Research Themes
Themes extracted and images generated with the OpenAI API; there may be inconsistencies.
Navigating Ethical AI Design
My research group explores the intersection of AI, ethics, and human-centered design to ensure that technology aligns with societal values. One important paper we focus on is [Mind the Gesture: Evaluating AI Sensitivity to Culturally Offensive Non-Verbal Gestures](https://arxiv.org/abs/2502.17710), which examines how AI systems can misinterpret cultural cues and foster harm. Additionally, we study [Mitigating Bias in RAG: Controlling the Embedder](https://arxiv.org/abs/2502.17390) to develop frameworks for AI that reduce biases. Our work is further informed by the findings in [Let Them Down Easy! Contextual Effects of LLM Guardrails on User Perceptions and Preferences](https://arxiv.org/abs/2506.00195), which highlights ways to enhance user trust through responsible AI design.
Exploring Narrative Dynamics
My research group explores the significance of narrative structures in understanding human experiences and AI interactions. We delve into the implications of narrative in technology through papers such as [Quantifying the narrative flow of imagined versus autobiographical stories](https://www.pnas.org/doi/10.1073/pnas.2211715119), which analyzes how stories impact our perceptions. Also noteworthy is [HEART-felt Narratives: Tracing Empathy and Narrative Style in Personal Stories with LLMs](https://arxiv.org/abs/2405.17633), which investigates emotional engagement in narratives produced by AI. Our exploration is further enhanced by analyzing [The Empirical Variability of Narrative Perceptions of Social Media Texts](https://aclanthology.org/2024.emnlp-main.1113/), shedding light on how narratives adapt across different platforms.
Enhancing AI Social Interaction
My research group explores the development of AI agents designed for social interactions, focusing on their capacity for social intelligence. A pivotal paper in this area is [SOTOPIA-S4: A User-Friendly System for Flexible, Customizable, and Large-Scale Social Simulation](https://arxiv.org/abs/2504.16122), which discusses how we can create more sophisticated simulations that mimic human behaviors. Another significant contribution is [Interactive Agents to Overcome Ambiguity in Software Engineering](https://arxiv.org/abs/2502.13069), addressing the potential of AI to support complex tasks. Moreover, we are interested in [Clever Hans or Neural Theory of Mind? Stress Testing Social Reasoning in Large Language Models](https://arxiv.org/abs/2305.14763), which evaluates how well AI systems can understand and replicate social contexts.
Advancing Conversational Technologies
My research group explores the innovative methodologies for enhancing conversational agents and their effectiveness in real-world applications. One significant work is [ProsocialDialog: A Prosocial Backbone for Conversational Agents](https://arxiv.org/abs/2205.12688), which proposes frameworks for promoting prosocial behavior within dialogue systems. We also focus on [SODA: Million-scale Dialogue Distillation with Social Commonsense Contextualization](https://arxiv.org/abs/2212.10465), aiming to capture the nuances of social reasoning in conversations. Critical insights are derived from [Aligning LLMs to Ask Good Questions: A Case Study in Clinical Reasoning](https://arxiv.org/abs/2502.14860), which explores the efficacy of AI in assisting healthcare conversations.