Focus on AI: Closing the Loop

University of Michigan School of Information faculty and PhD students are advancing the field of artificial intelligence through innovative research and impactful contributions. Here are some of their recent publications.

Publications 

The Imitation of Intimacy: Comparing Satisfaction in Intimate Human and AI Companion Relationships

HRI '26: Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction, March 2026

Annette Masterson, Xin Ye, Yiyang Li, Lionel Robert

The rapid proliferation of Large Language Models (LLMs) has enabled artificial agents to foster deep emotional bonds, yet the comparability of these AI relationships to human norms remains underexplored. As HRI researchers increasingly integrate LLMs into embodied platforms, understanding the nature of these bonds is imperative for responsible design. This study investigates whether relationships with LLM-driven AI companions can rival the satisfaction of human connections and if the mechanism of intimacy is equally critical. Through a comparative survey of 150 participants stratified across in-person, long-distance, and LLM companion relationships, we illuminate that digital bonds can yield satisfaction levels comparable to human partnerships, with intimacy serving as a predictive factor. These findings challenge the assumption that AI relationships are inherently unsatisfactory and identify intimacy as a design metric for social robots, providing a protocol for integrating LLM companions into embodied relational agents.

The Roles of Fairness and Effectiveness in Promoting Legitimacy and Cooperation with Security Robotic Authority

HRI '26: Proceedings of the 21st ACM/IEEE International Conference on Human-Robot Interaction, March 2026

Xin Ye, Lionel Robert

Security robots increasingly assume authoritative roles, but the underlying mechanisms for why humans cooperate with them is not well understood. This study proposed and tested a cooperation model based on legitimacy theory, focusing on how distributive fairness (outcome equity) and interactional fairness (treatment equity) influence robot legitimacy and cooperation. Using a 2 × 2 online video-based experiment with 372 U.S. participants, the authors found that both fairness types promote cooperation through value alignment, with a non-significant path through obligation to obey; meanwhile, perceived effectiveness was strongly associated with both value alignment and obligation to obey. These findings extend legitimacy theory to human–robot interaction in a U.S. context, emphasizing fairness and perceived effectiveness as key to fostering cooperation and informing ethical robot design.

AI Evaluation and the Standards Metaphor

Yale Journal of Law & Technology, March 2026 

Amina A. Abdu, Abigail Z. Jacobs

Significant attention has been devoted to the question of how best to govern artificial intelligence (AI). In addition to legislation, many policy proposals focus on extra-legal regulatory instruments. Notably, AI evaluations provide a particularly attractive solution, imposing seemingly neutral measurements across the widespread contexts in which AI operates. Because AI evaluations are driven by a wide range of actors, their adoption as a governance tool is shifting power in AI policymaking. In particular, the companies that create AI are also key players in designing and marketing AI evaluations. This Essay examines how large technology companies and government actors conceptualize self-regulation by technology companies as a legitimate policy intervention. We note that AI evaluations are often described using the language of standards, another more established soft law regulatory instrument. Drawing on the history of standards, we discuss how AI companies leverage the metaphor of standards to describe benchmarks and evaluations in order to legitimate corporate expertise. We then examine the implications of this metaphor, describing where it is useful in the context of AI and where it obscures important policy decisions.

Digital Twins for Just-in-Time Adaptive Interventions (JITAIs): Framework for Optimizing and Continually Improving JITAIs

Journal of Medical Internet Research, March 2026 

Asim H Gazi, Daiqi Gao, Susobhan Ghosh, Ziping Xu, Anna L Trella, Predrag Klasnja, Susan A Murphy

In the context of digital health, just-in-time adaptive interventions (JITAIs) are nascent precision medicine systems that can extend personalized health care support to everyday life. A challenge in designing JITAIs is that personalized support often involves sophisticated decision-making algorithms. These decision-making algorithms can require numerous nontrivial design decisions that must be made between successive JITAI deployments (eg, hyperparameter selection for an artificial intelligence algorithm). Making design decisions between deployments—rather than during deployment—ensures intervention fidelity and enhances the ability to replicate results. Yet, each deployment can be costly, precluding the use of A/B testing for every design decision. How should design decisions be made strategically between JITAI deployments? This paper introduces “digital twins for just-in-time adaptive interventions (JITAI-Twins)” to address this question. JITAI-Twins are “digital twins of a subpopulation” (term used in the 2023 National Academies workshop proceedings on digital twins). JITAI-Twins are used to virtually simulate the potential outcomes of a JITAI’s design decisions for an upcoming deployment. Based on simulation results, design decisions are made for the deployed JITAI. To continually improve the JITAI, data collected during deployment are used to update the JITAI-Twin—and this bidirectional feedback between deployments and simulation environments continues. JITAI-Twins are thus “fit-for-purpose” (term used in the National Academies 2024 consensus report on digital twins) instantiations of the digital twin concept. In this paper, we elucidate the specifics and design process of JITAI-Twins, with examples of prior use in clinical settings. JITAI-Twins highlight continuity over the course of a JITAI’s optimization and continual improvement, emphasizing the need for bidirectional feedback between versions of a simulation environment and a JITAI’s deployments.

Closing the Loop: An Instructor-in-the-Loop AI Assistance System for Supporting Student Help-Seeking in Programming Education

SIGCSE TS 2026: Proceedings of the 57th ACM Technical Symposium on Computer Science Education V.1, February 2026 

*Best Paper Award

Tung Phung, Heeryung Choi, Mengyan Wu, Christopher Brooks, Sumit Gulwani, Adish Singla

Timely and high-quality feedback is essential for effective learning in programming courses; yet, providing such support at scale remains a challenge. While AI-based systems offer scalable and immediate help, their responses can occasionally be inaccurate or insufficient. Human instructors, in contrast, may bring more valuable expertise but are limited in time and availability. To address these limitations, we present a hybrid help framework that integrates AI-generated hints with an escalation mechanism, allowing students to request feedback from instructors when AI support falls short. This design leverages the strengths of AI for scale and responsiveness while reserving instructor effort for moments of greatest need. We deployed this tool in a data science programming course with 82 students. We observe that out of the total 673 AI-generated hints, students rated 146 (22%) as unhelpful. Among those, only 16 (11%) of the cases were escalated to the instructors. A qualitative investigation of instructor responses showed that those feedback instances were incorrect or insufficient roughly half of the time. This finding suggests that when AI support fails, even instructors with expertise may need to pay greater attention to avoid making mistakes. We will publicly release the tool for broader adoption and enable further studies in other classrooms. Our work contributes a practical approach to scaling high-quality support and informs future efforts to effectively integrate AI and humans in education.

CA–CI: Integrating Contextual Integrity and the Capabilities Approach for Dignity Considerations in AI Governance

IEEE Explore, January 2026 

Kat Roemmich, Kirsten Martin, Florian Schaub

Capabilities approach -contextual integrity (CA–CI) extends contextual integrity through the integration of dignity thresholds from the capabilities approach and the specification of purpose as a constitutive parameter. We demonstrate how CA–CI can operationalize the EU AI Act's fundamental rights impact assessments, harm thresholds, and anticipatory governance.

Pre-prints, Working Papers, Articles, Workshops and Talks

Editrail: Understanding AI Usage by Visualizing Student-AI Interaction in Code

arXiv, January 2026

Ashley Ge Zhang, Yan-Ru Jhou, Yinuo Yang, Shamita Rao, Maryam Arab, Yan Chen, Steve Oney

Programming instructors have diverse philosophies about integrating generative AI into their classes. Some encourage students to use AI, while others restrict or forbid it. Regardless of their approach, all instructors benefit from understanding how their students actually use AI while writing code. Such insight helps instructors assess whether AI use aligns with their pedagogical goals, enables timely intervention when they find unproductive usage patterns, and establishes effective policies for AI use. However, our survey with programming instructors found that many instructors lack visibility into how students use AI in their code-writing processes. To address this challenge, we introduce Editrail, an interactive system that enables instructors to track students' AI usage, create personalized assessments, and provide timely interventions, all within the workflow of monitoring coding histories. We found that Editrail enables instructors to detect AI use that conflicts with pedagogical goals accurately and to determine when and which students require intervention.

RELATED

Keep up with research on AI from UMSI experts by subscribing to our free research roundup newsletter!