top of page

Malihe Alikhani
 

I study AI systems in context across language, embodiment, and society, with a focus on enhancing communication, decision making, and knowledge sharing across disciplines and populations. My work bridges academia, applied AI research, and public policy, shaped by my roles as an Assistant Professor of Computer Science at Northeastern University and a Resident Visiting Fellow at the Brookings Institution, to help ensure that AI serves society responsibly and effectively.
 

I lead the Contextual AI Lab, where we build embodied systems that support productive interactions between humans and machines. By integrating insights from cognitive and social sciences with machine learning, we develop models that capture the richness of human interpretation and support collaborative meaning construction.
 

I have served as a science and technology advisor in Congress and collaborated with the United Nations in Africa on initiatives using AI to support anti-violence efforts and promote equitable access to education and health.
 

I also serve as Ethics Chair of the Association for Computational Linguistics, a member of the U.S. Technology Policy Committee of the Association for Computing Machinery, and on the Editorial Board of the Journal of Artificial Intelligence Research.
 


email: m.alikhani AT northeastern.edu

20260114_Malihe_007_edited.jpg

Recent Research Themes

Alignment and Evaluation for Reliable Human AI Collaboration
 

I develop evaluation and adaptation frameworks for AI systems embedded in human decision workflows. This work studies when AI systems are reliable, for whom they are reliable, and how their uncertainty should be communicated in interactive settings. Across this line of work, I model demographic variability, distribution shift, override capacity, calibration, and appropriate reliance in human AI systems. The broader goal is to move AI evaluation beyond static benchmarks toward methods that capture how people actually interpret, contest, and depend on AI outputs in consequential contexts.
 

Selected publications

  1. HumBEL: A Human in the Loop Approach for Evaluating Demographic Factors of Language Models in Human Machine Conversations, Anthony Sicilia, Malihe Alikhani, EACL 2024.

  2. Adaptive Platt Scaling with Causal Interpretations for Self Reflective Language Model Uncertainty Estimates, Anthony Sicilia, Malihe Alikhani, Findings of ACL: EMNLP 2025.

  3. An Active Learning Framework for Inclusive Generation by Large Language Models, Sabit Hassan, Anthony B. Sicilia, Malihe Alikhani, COLING 2025.

  4. Active Learning for Robust and Representative LLM Generation in Safety Critical Scenarios, Sabit Hassan, Anthony Sicilia, Malihe Alikhani, CustomNLP4U 2024. Best Paper Award.

  5. PAC Bayesian Domain Adaptation Bounds for Multiclass Learners, Anthony Sicilia, Katherine Atwell, Malihe Alikhani, Seong Jae Hwang, UAI 2022. Best Paper Award.

  6. Better Slow than Sorry: Introducing Positive Friction for Reliable Dialogue Systems, Mert Inan, Anthony Sicilia, Suvodip Dey, Vardhan Dongre, Tejas Srinivasan, Jesse Thomason, Gökhan Tür, Dilek Hakkani Tür, Malihe Alikhani, TACL 2025.

 

Multimodal Communication with AI: Safety and Productivity in Health and Education

Developing multimodal AI systems, spanning robots, virtual agents, and interactive platforms, alongside the datasets and benchmarks necessary for their deployment in embodied and socially embedded environments. This work evaluates system robustness, demographic variability, and safety across real-world classroom, healthcare, and assistive settings.


Selected Papers: 
Better slow than sorry: Introducing positive friction for reliable dialogue systems, Mert Inan, Anthony Sicilia, Suvodip Dey, Vardhan Dongre, Tejas Srinivasan, Jesse Thomason, Gökhan Tür, Dilek Hakkani-Tür, Malihe Alikhani, TACL 2025. 

SODA: Million-scale dialogue distillation with social commonsense contextualization, Hyunwoo Kim, Jack Hessel, Liwei Jiang, Peter West, Ximing Lu, Youngjae Yu, Pei Zhou, Ronan Bras, Malihe Alikhani, Gunhee Kim, Maarten Sap, Yejin Choi, Proceedings of the 2024Conference on Empirical Methods in Natural Language Processing, 2024. Best Paper Award

SiLVERScore: Semantically-Aware Embeddings for Sign Language Generation Evaluation, Saki Imai, Mert Inan, Anthony B Sicilia, Malihe Alikhani, Proceedings of the 15th International Conference on Recent Advances in Natural Language Processing-Natural Language Processing in the Generative AI Era, 2025.

Studying and mitigating biases in sign language understanding models, Katherine Atwell, Danielle Bragg, Malihe Alikhani, Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 2024.

SignAlignLM: Integrating multimodal sign language processing into large language models, Mert Inan, Anthony Sicilia, Malihe Alikhani, Findings of the Association for Computational Linguistics: ACL 2025, 2025.

Identifying & Interactively Refining Ambiguous User Goals for Data Visualization Code Generation, Mert Inan, Anthony Sicilia, Alex Xie, Saujas Vaduguru, Daniel Fried, Malihe Alikhani, Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, 2025.

Cross-modal Coherence Modeling for Caption Generation, Malihe Alikhani, Piyush Sharma, Shengjie Li, Rad Soricut, Matthew Stone, Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, 2020.

MedNgage: A dataset for understanding engagement in patient-nurse conversations, Yan Wang, Heidi Donovan, Sabit Hassan, Malihe Alikhani, Findings of the Association for Computational Linguistics: ACL 2023.

Including Signed Languages in Natural Language Processing, Kayo Yin, Amit Moryossef, Julie Hochgesang, Yoav Goldberg, and Malihe Alikhani, The Joint Conference of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing, 2021. Best  Theme Paper Award

Uncertainty, Belief, and Influence in AI-Mediated Communication
 

I study AI systems as communicative agents that participate in belief formation, social reasoning, persuasion, and decision making. This work develops computational methods for modeling uncertainty, belief revision, theory of mind, sycophancy, preference strength, and social influence in dialogue. Rather than treating model responses as isolated outputs, this thrust examines how AI systems shape human interpretation, trust, perceived bias, and decision trajectories over time. This line of work connects technical questions in pragmatics and uncertainty modeling with broader societal concerns around manipulation, polarization, misinformation, and responsible AI deployment.
 

Selected publications

  1. Evaluating Theory of an Uncertain Mind: Predicting the Uncertain Beliefs of Others in Conversation Forecasting, Anthony Sicilia and Malihe Alikhani, NAACL 2025.

  2. Deal, or No Deal or Who Knows? Forecasting Uncertainty in Conversations Using Large Language Models, Anthony Sicilia, Hyunwoo Kim, Khyathi Raghavi Chandu, Malihe Alikhani, Jack Hessel, ACL 2024.

  3. Accounting for Sycophancy in Language Model Uncertainty Estimation, Anthony Sicilia, Mert Inan, Malihe Alikhani, NAACL 2025.

  4. BASIL: Bayesian Assessment of Sycophancy in LLMs, Katherine Atwell, Pedram Heydari, Anthony Sicilia, Malihe Alikhani, FAccT 2026.

  5. MixDPO: Modeling Preference Strength for Pluralistic Alignment, Saki Imai, Pedram Heydari, Anthony Sicilia, Asteria Kaeberlein, Katherine Atwell, Malihe Alikhani, arXiv preprint, 2026.

  6. Eliciting Uncertainty in Chain of Thought to Mitigate Bias Against Forecasting Harmful User Behaviors, Anthony Sicilia, Malihe Alikhani, NLP for Positive Impact 2024. Best Paper Award.

  7. How LLMs Influence Perceived Bias in Journalism, Asteria Kaeberlein, Malihe Alikhani, RANLP 2025.

  8. How People Talk About Each Other: Modeling Generalized Intergroup Bias and Emotion, Venkata Subrahmanyan Govindarajan, Katherine Atwell, Barea Sinno, Malihe Alikhani, David I. Beaver, Junyi Jessy Li, EACL 2023.

  9. Political Ideology and Polarization: A Multi-Dimensional Approach, Barea Sinno, Bernardo Oviedo, Katherine Atwell, Malihe Alikhani, Junyi Jessy Li, NAACL 2022.

  10. How to Ask for Donations? Learning User Specific Persuasive Dialogue Policies through Online Interactions, Nhat Tran, Malihe Alikhani, Diane Litman, UMAP 2022. Best Paper Award.

  11. Studying the Effect of Moderator Biases on the Diversity of Online Discussions: A Computational Cross-Linguistic Study, Sabit Hassan, Katherine J. Atwell, Malihe Alikhani, Cognitive Science Society 2022.

 

bottom of page