arXiv subjects

Mark Riedl

Publications and source records attributed to Mark Riedl.

At least 19 recordsLinked to original sources

arXiv2026

No One Wins in Nuclear War: A Social Simulation of Military Decision-making

WOPR is a social-simulation environment for studying how organizations make high-stakes decisions, built on a deterministic, replay-validated rules engine and using wargames as the vehicle. We instantiate it first with the published card game Nuclear War, traced against its published rules. We start with military decision-making because of its safety implications and because it needs further study, but the design is not specific to it: the decision-point contract that exposes the engine to agents is reusable across verifiable rule systems. Existing social-simulation work emphasizes persona fidelity and synthetic opinion, but lacks a verifiable rules engine with replay-checkable mechanics and private-channel negotiation. WOPR supplies that engine, and its contract makes every strategic choice an explicit agent decision. The method is agnostic to social-simulation frameworks; we adopt Concordia as the default harness for driving the game. On the same engine, WOPR layers a four-rung press ladder from silence to private single-recipient channels with structured commitments, and instantiates each faction as a collective command-and-control system rather than a single agent. We make all code, example configurations, and replay data publicly available at https://github.com/eilab-gt/wopr.

cs.CY↗

arXiv2026

Role Steering of Language Models for Social Simulations

Social simulations built from language-model agents need role-conditioned behavior that can be checked before agents are placed into a simulated population. We introduce an activation-steering screening workflow for role-conditioned agents: define a role profile, extract a role-specific direction, sweep four steering coefficients, evaluate role-profile alignment, and pass or flag each candidate configuration. On OLMo-3-7B-Instruct, we apply the workflow to a mixed 275-role inventory with 228 role-agnostic questions, GPT-4.1-mini prompted role references, and GPT-4.1-mini judges. Role-specific directions receive higher judged role-profile alignment than an assistant-axis directional control from prior persona-vector work, with mean overall scores of 63.2 versus 41.1 across the tested grid. They also preserve high lexical diversity, while the control drops sharply at larger coefficients. The role-level screen is the main practical output: most roles improve as steering increases, but 38 roles decline across all six measured dimensions, showing why simulation builders should choose coefficients per role rather than deploy a uniform high-strength setting. We make our code and evaluation artifacts available at https://anonymous.4open.science/r/anonymous-research-code-5F03/.

cs.CL↗

arXiv2026

Concept Catalyst: Exploring Scrutable Interfaces to Structure K-12 Teacher Interactions with Generative AI

Purpose: This paper explores how to align AI-based tools with teachers' classroom needs by using scrutable interfaces -- interfaces that link an easily manipulable knowledge representation to an underlying AI model, so users can change the system's outputs without understanding its details. It provides an in-depth discussion and example of a scrutable interface that structures teachers' interactions with generative AI. This study aims to expand how and where scrutable interfaces are used in AI-based tools to support teachers, who have not been historically targeted in the design of scrutable systems. Design/Methodology/Approach: This paper presents the design and evaluation of Concept Catalyst, an AI-based tool with a scrutable interface, created to support teachers' reflection while using generative AI for curriculum development. It presents the findings from an exploratory study using Wizard-of-Oz testing with middle and high school engineering teachers, resulting in 10 depth interviews lasting 55 minutes on average. Screen/audio recordings and the classroom content teachers produced during the session were also collected. Findings: The paper provides empirical insights about how scrutable interfaces can positively structure teachers' interactions with generative AI models when creating classroom content. Findings suggest that scrutable interfaces can help teachers reflect on their teaching practices while improving efficacy, efficiency, and motivation when using AI. What is original/value of the paper: This paper explores an identified need to support teachers' classroom practices and needs when using generative AI. It extends the consideration of scrutable interfaces in two ways: to support teachers as users (not just students) and to structure interactions with generative AI models.

cs.HC↗

arXiv2026

Drawing Out Legal Risks: Co-Designing with Lawyers to Predict and Manage Legal Uncertainties of Medical AI Tools

While there's optimism around medical AI tools due to their abilities to adapt from user-to-user and across environments, these new abilities complicate how people and organizations are able to predict and manage risk based on existing laws and regulations. Lawyers are trained to identify potential legal outcomes, but they lack technical AI knowledge, making it difficult to translate their expertise to creators and users of AI tools. We contribute insights from our co-design process with U.S. lawyers to identify and translate ways to predict and manage risks of medical AI tools. We present the visualizations we developed through two years of cross-disciplinary efforts and thereby illustrate our findings about how legal risks are determined and our strategies for people and organizations to predict and manage these risks. We offer insights about leveraging lawyers' expertise to understand, predict, and manage legal risks.

cs.HC↗

arXiv2026

Capability Provenance in Language Models: A Case Study in Social Reasoning

We use training-data attribution as an interpretable tool for capability discovery, mapping which regions of the pretraining corpus support social-reasoning versus STEM-reasoning in OLMo3-7B. Training-data attribution measures how strongly each training document influences a model's predictions on a benchmark, but document-level scores are too noisy to identify which corpus regions support which capabilities. We compute gradient-based attribution (TrackStar via Bergson) over a working set drawn from the de-duplicated Dolma3 mix, aggregate influence across WebOrganizer's 24-format x 24-topic taxonomy (576 bins), and contrast benchmark pairs in a 2x2 design that varies domain (social vs. STEM) and capability type (reasoning vs. knowledge): SocialIQA and MMLU Social Sciences against ARC-Challenge and MMLU STEM. Social and STEM reasoning draw on qualitatively distinct corpus regions, and the contrast is sharper at the reasoning level than at the knowledge level. Targeted machine unlearning provides partial causal validation: forgetting high-attribution topics (e.g., Literature for SocialIQA) degrades the aligned benchmark more than within-topic random baselines. We validate on other open-data model, Comma v0.1 7B-2T (Common Pile) and DCLM-Baseline-7B (DataComp-LM): causal selectivity holds on both models, while the provenance map is ecosystem-specific. We open-source all code, data artifacts, influence scores, and checkpoints at https://github.com/eilab-gt/capabilibara and https://huggingface.co/HCAI-Lab.

cs.CL↗

arXiv2026

Explainable Model Routing for Agentic Workflows

Modern agentic workflows decompose complex tasks into specialized subtasks and route them to diverse models to minimize cost without sacrificing quality. However, current routing architectures focus exclusively on performance optimization, leaving underlying trade-offs between model capability and cost unrecorded. Without clear rationale, developers cannot distinguish between intelligent efficiency -- using specialized models for appropriate tasks -- and latent failures caused by budget-driven model selection. We present Topaz, a framework that introduces formal auditability to agentic routing. Topaz replaces silent model assignments with an inherently interpretable router that incorporates three components: (i) skill-based profiling that synthesizes performance across diverse benchmarks into granular capability profiles (ii) fully traceable routing algorithms that utilize budget-based and multi-objective optimization to produce clear traces of how skill-match scores were weighed against costs, and (iii) developer-facing explanations that translate these traces into natural language, allowing users to audit system logic and iteratively tune the cost-quality tradeoff. By making routing decisions interpretable, Topaz enables users to understand, trust, and meaningfully steer routed agentic systems.

cs.AI↗

arXiv2026

Evaluating Actionability in Explainable AI

A core assumption of Explainable AI (XAI) is that explanations are useful to users -- that is, users will do something with the explanations. Prior work, however, does not clearly connect the information provided in explanations to user actions to evaluate effectiveness. In this paper, we articulate this connection. We conducted a formative study through 14 interviews with end users in education and medicine. We contribute a catalog of information and associated actions. Our catalog maps 12 categories of information that participants described relying on to take 60 different actions. We show how AI Creators can use the catalog's specificity and breadth to articulate how they expect information in their explanations to lead to user actions and test their assumptions. We use an exemplar XAI system to illustrate this approach. We conclude by discussing how our catalog expands the design space for XAI systems to support actionability.

cs.HC↗

arXiv2025

World Model Robustness via Surprise Recognition

AI systems deployed in the real world must contend with distractions and out-of-distribution (OOD) noise that can destabilize their policies and lead to unsafe behavior. While robust training can reduce sensitivity to some forms of noise, it is infeasible to anticipate all possible OOD conditions. To mitigate this issue, we develop an algorithm that leverages a world model's inherent measure of surprise to reduce the impact of noise in world model--based reinforcement learning agents. We introduce both multi-representation and single-representation rejection sampling, enabling robustness to settings with multiple faulty sensors or a single faulty sensor. While the introduction of noise typically degrades agent performance, we show that our techniques preserve performance relative to baselines under varying types and levels of noise across multiple environments within self-driving simulation domains (CARLA and Safety Gymnasium). Furthermore, we demonstrate that our methods enhance the stability of two state-of-the-art world models with markedly different underlying architectures: Cosmos and DreamerV3. Together, these results highlight the robustness of our approach across world modeling domains. We release our code at https://github.com/Bluefin-Tuna/WISER .

cs.LG↗

arXiv2025

Shall We Play a Game? Language Models for Open-ended Wargames

LLM-based social simulations can make a generated transcript look like a single behavioral signal, but the model behind that transcript may be doing several different jobs: choosing what an actor says or does, deciding what happens after an action, or both. The difference matters especially in open-ended wargames, where models are prized for handling unusual actions and ambiguous consequences. We report a scoping review of 223 de-duplicated AI-in-wargames and strategic-simulation papers retrieved through May 1, 2026, describing each simulation by its model-control profile: whether the language model has open-ended control over player actions, adjudication, or both. Only 20 of 223 studies (~9%) give language models both roles. Before treating LM outputs as social simulations, researchers need to know how much creative control the model has over actions and consequences. For open-ended simulations, fidelity depends not only on whether agents behave plausibly, but also on whether language models can reliably act as adjudicators or world models.

cs.AI↗

arXiv2025

Do Language Models Agree with Human Perceptions of Suspense in Stories?

Suspense is an affective response to narrative text that is believed to involve complex cognitive processes in humans. Several psychological models have been developed to describe this phenomenon and the circumstances under which text might trigger it. We replicate four seminal psychological studies of human perceptions of suspense, substituting human responses with those of different open-weight and closed-source LMs. We conclude that while LMs can distinguish whether a text is intended to induce suspense in people, LMs cannot accurately estimate the relative amount of suspense within a text sequence as compared to human judgments, nor can LMs properly capture the human perception for the rise and fall of suspense across multiple text segments. We probe the abilities of LM suspense understanding by adversarially permuting the story text to identify what cause human and LM perceptions of suspense to diverge. We conclude that, while LMs can superficially identify and track certain facets of suspense, they do not process suspense in the same way as human readers.

cs.CL↗

arXiv2025

Implications of Current Litigation on the Design of AI Systems for Healthcare Delivery

Many calls for explainable AI (XAI) systems in medicine are tied to a desire for AI accountability--accounting for, mitigating, and ultimately preventing harms from AI systems. Because XAI systems provide human-understandable explanations for their output, they are often viewed as a primary path to prevent harms to patients. However, when harm occurs, laws, policies, and regulations also shape AI accountability by impacting how harmed individuals can obtain recourse. Current approaches to XAI explore physicians' medical and relational needs to counter harms to patients, but there is a need to understand how XAI systems should account for the legal considerations of those impacted. We conduct an analysis of 31 legal cases and reported harms to identify patterns around how AI systems impact patient care. Our findings reflect how patients' medical care relies on a complex web of stakeholders--physicians, state health departments, health insurers, care facilities, among others--and many AI systems deployed across their healthcare delivery negatively impact their care. In response, patients have had no option but to seek legal recourse for harms. We shift the frame from physician-centered to patient-centered accountability approaches by describing how lawyers and technologists need to recognize and address where AI harms happen. We present paths for preventing or countering harm (1) by changing liability structures to reflect the role of many stakeholders in shaping how AI systems impact patient care; and (2) by designing XAI systems that can help advocates, such as legal representatives, who provide critical legal expertise and practically support recourse for patients.

cs.HC↗

arXiv2025

Understanding the Impact of Physicians' Legal Considerations on XAI Systems

Physicians are--and feel--ethically, professionally, and legally responsible for patient outcomes, buffering patients from harmful AI determinations from medical AI systems. Many have called for explainable AI (XAI) systems to help physicians incorporate medical AI recommendations into their workflows in a way that reduces the potential of harms to patients. While prior work has demonstrated how physicians' legal concerns impact their medical decision making, little work has explored how XAI systems should be designed in light of these concerns. In this study, we conducted interviews with 10 physicians to understand where and how they anticipate errors that may occur with a medical AI system and how these anticipated errors connect to their legal concerns. In our study, physicians anticipated risks associated with using an AI system for patient care, but voiced unknowns around how their legal risk mitigation strategies may change given a new technical system. Based on these findings, we describe the implications for designing XAI systems that can address physicians' legal concerns. Specifically, we identify the need to provide AI recommendations alongside contextual information that guides their risk mitigation strategies, including how non-legally related aspects of their systems, such as medical documentation and auditing requests, might be incorporated into a legal case.

cs.HC↗

arXiv2025

Legally-Informed Explainable AI

Explanations for artificial intelligence (AI) systems are intended to support the people who are impacted by AI systems in high-stakes decision-making environments, such as doctors, patients, teachers, students, housing applicants, and many others. To protect people and support the responsible development of AI, explanations need to be actionable--helping people take pragmatic action in response to an AI system--and contestable--enabling people to push back against an AI system and its determinations. For many high-stakes domains, such as healthcare, education, and finance, the sociotechnical environment includes significant legal implications that impact how people use AI explanations. For example, physicians who use AI decision support systems may need information on how accepting or rejecting an AI determination will protect them from lawsuits or help them advocate for their patients. In this paper, we make the case for Legally-Informed Explainable AI, responding to the need to integrate and design for legal considerations when creating AI explanations. We describe three stakeholder groups with different informational and actionability needs, and provide practical recommendations to tackle design challenges around the design of explainable AI systems that incorporate legal considerations.

cs.HC↗

arXiv2025

The Goofus & Gallant Story Corpus for Practical Value Alignment

Values or principles are key elements of human society that influence people to behave and function according to an accepted standard set of social rules to maintain social order. As AI systems are becoming ubiquitous in human society, it is a major concern that they could violate these norms or values and potentially cause harm. Thus, to prevent intentional or unintentional harm, AI systems are expected to take actions that align with these principles. Training systems to exhibit this type of behavior is difficult and often requires a specialized dataset. This work presents a multi-modal dataset illustrating normative and non-normative behavior in real-life situations described through natural language and artistic images. This training set contains curated sets of images that are designed to teach young children about social principles. We argue that this is an ideal dataset to use for training socially normative agents given this fact.

cs.AI↗

arXiv2024

Multi-Attribute Constraint Satisfaction via Language Model Rewriting

Obeying precise constraints on top of multiple external attributes is a common computational problem underlying seemingly different domains, from controlled text generation to protein engineering. Existing language model (LM) controllability methods for multi-attribute constraint satisfaction often rely on specialized architectures or gradient-based classifiers, limiting their flexibility to work with arbitrary black-box evaluators and pretrained models. Current general-purpose large language models, while capable, cannot achieve fine-grained multi-attribute control over external attributes. Thus, we create Multi-Attribute Constraint Satisfaction (MACS), a generalized method capable of finetuning language models on any sequential domain to satisfy user-specified constraints on multiple external real-value attributes. Our method trains LMs as editors by sampling diverse multi-attribute edit pairs from an initial set of paraphrased outputs. During inference, LM iteratively improves upon its previous solution to satisfy constraints for all attributes by leveraging our designed constraint satisfaction reward. We additionally experiment with reward-weighted behavior cloning to further improve the constraint satisfaction rate of LMs. To evaluate our approach, we present a new Fine-grained Constraint Satisfaction (FineCS) benchmark, featuring two challenging tasks: (1) Text Style Transfer, where the goal is to simultaneously modify the sentiment and complexity of reviews, and (2) Protein Design, focusing on modulating fluorescence and stability of Green Fluorescent Proteins (GFP). Our empirical results show that MACS achieves the highest threshold satisfaction in both FineCS tasks, outperforming strong domain-specific baselines. Our work opens new avenues for generalized and real-value multi-attribute control, with implications for diverse applications spanning NLP and bioinformatics.

cs.AI↗

arXiv2024

Recognizing Lawyers as AI Creators and Intermediaries in Contestability

Laws play a key role in the complex socio-technical system impacting contestability: they create the regulations shaping the way AI systems are designed, evaluated, and used. Despite their role in the AI value chain, lawyers' impact on contestability has gone largely unrecognized in the design of AI systems. In this paper, we highlight two main roles lawyers play that impact contestability: (1) as AI Creators because the regulations they create shape the design and evaluation of AI systems before they are deployed; and (2) as Intermediaries because they interpret regulations when harm occurs, navigating the gap between stakeholders, instutions, and harmful outcomes. We use these two roles to illuminate new opportunities and challenges for including lawyers in the design of AI systems, contributing a significant first step in practical recommendations to amplify the power to contest systems through cross-disciplinary design.

cs.HC↗

arXiv2024

Making Large Language Models into World Models with Precondition and Effect Knowledge

World models, which encapsulate the dynamics of how actions affect environments, are foundational to the functioning of intelligent agents. In this work, we explore the potential of Large Language Models (LLMs) to operate as world models. Although LLMs are not inherently designed to model real-world dynamics, we show that they can be induced to perform two critical world model functions: determining the applicability of an action based on a given world state, and predicting the resulting world state upon action execution. This is achieved by fine-tuning two separate LLMs-one for precondition prediction and another for effect prediction-while leveraging synthetic data generation techniques. Through human-participant studies, we validate that the precondition and effect knowledge generated by our models aligns with human understanding of world dynamics. We also analyze the extent to which the world model trained on our synthetic data results in an inferred state space that supports the creation of action chains, a necessary property for planning.

cs.CL↗

arXiv2024

Beyond Following: Mixing Active Initiative into Computational Creativity

Generative Artificial Intelligence (AI) encounters limitations in efficiency and fairness within the realm of Procedural Content Generation (PCG) when human creators solely drive and bear responsibility for the generative process. Alternative setups, such as Mixed-Initiative Co-Creative (MI-CC) systems, exhibited their promise. Still, the potential of an active mixed initiative, where AI takes a role beyond following, is understudied. This work investigates the influence of the adaptive ability of an active and learning AI agent on creators' expectancy of creative responsibilities in an MI-CC setting. We built and studied a system that employs reinforcement learning (RL) methods to learn the creative responsibility preferences of a human user during online interactions. Situated in story co-creation, we develop a Multi-armed-bandit agent that learns from the human creator, updates its collaborative decision-making belief, and switches between its capabilities during an MI-CC experience. With 39 participants joining a human subject study, Our developed system's learning capabilities are well recognized compared to the non-learning ablation, corresponding to a significant increase in overall satisfaction with the MI-CC experience. These findings indicate a robust association between effective MI-CC collaborative interactions, particularly the implementation of proactive AI initiatives, and deepened understanding among all participants.

cs.HC↗