Center for Ethical AI: Projects

Projects

• AI-Powered Adaptive Story-Based Authentication for Immersive Environments

  This research integrates LLM-based question generation to evaluate security improvements through theoretical analysis of entropy and shouldersurfing resistance.

• Applied Ethical AI for Protection and Resilience of Modern Power Grids

  This research focuses on how ethical principles can be defined, evaluated, and put into practice when AI is used to protect critical infrastructure and enhance grid resilience.

• Beyond Survey Preferences: Ethical Risks of Emotional Reliance in Adolescents with Extensive AI Relationships

  This research evaluates how generative AI chatbots shape adolescents’ emotional bonding and reliance when youth already treat AI as a social partner.

• Bias, Accuracy, and Cultural Representation in AI-Generated Language: An Ethical AI Study in Language Education

  This interdisciplinary project investigates bias, linguistic accuracy, and cultural representation in AI-generated Arabic within university-level language instruction.

• Building Cross-Domain Knowledge on AI Risk Assessment and Mitigation Frameworks

  This interdisciplinary research project investigates how ethical AI risk assessment frameworks perform across diverse disciplinary contexts and develop evidence-based innovations to improve their effectiveness.

• Procedural vs. Narrative Agency: Comparative Analysis of State Records and Survivor Graffiti using Offline LLMs

  Using offline LLMs, this research examines the contrasting narratives of intimate partner violence (IPV) through two distinct textual sources: "procedural" narratives in police reports, which represent the state's construction of victimhood, and "sanctuary" narratives in restroom graffiti, which capture victims' own voices and self-representation.

• “Don’t look at me!”: How managers use artificial intelligence to shift responsibility

  This study aims examines whether managers are more likely to emphasize AI involvement when they anticipate blame for unfavorable organizational decisions? 

• Ethical AI for Developing Countries: India and the ZOHO model

  This research examines how developing countries avoid being left behind in the race to develop AI.

• Ethical Boundaries of AI-Mediated Relationship Advice: An Interdisciplinary Study of Trust, Autonomy, and Responsibility

  This projects examines AI applications in relationship advice, integrating perspectives from psychology, computer science, ethics, and public health.

• Ethical Saliency and Choice of AI- Versus Human-Generated Graphs

  This study focuses on the choice of graph between Human- and AI-generated content.

• Ethically Aligned AI for Point-of-Care CE-LIF Detection of Brain Hemorrhage and Neurovascular Injury

  This study integrates machine learning and ethical generative AI into a rapid point-of-care blood test based on capillary electrophoresis with laser-induced fluorescence (CE-LIF) to detect intracranial hemorrhage and neurovascular injury.

• Generative AI Impacts, Framing, and Policy Support

  This study involves a series of survey experiments that vary the frames around the costs and benefits of generative AI systems and assess how these frames affect perceptions of the ethicality of generative AI, perceptions of risk and benefits, and support for public policy interventions.

• Human–AI Collaboration and The Future of Work

  This interdisciplinary research focuses on how varying degrees of algorithmic autonomy shape managerial responsibility, accountability, and fairness

• Impacts of Generative AI on Science Literacy

  This project aims to understand how student interaction with Generative AI (GenAI) platforms influences science literacy.

• Improving Evidence Review in Public Defense with AI Tools

  This study uses custom Large Language Models (LLMs) and Automatic Speech Recognition (ASR) models to create a pipeline for analyzing and comparing police bodycam footage to police reports.

• Measuring Semantic Fidelity in LLM-Based Privacy Policy Interpretation for Older Adults

  This study aims to use large language models (LLMs) to automatically simplify complex privacy policies into plain language, increasing their accessibility and interpretability for older adults who use IoT (Internet of Things) healthcare devices.

• Supervised differential network analysis for interpretable Alzheimer’s detection

  This project involved expanding research on supervised machine learning methods for functional brain mapping. Rather than using a static map, the investigators used machine learning to identify and group brain regions based on their contribution to a specific classification task, focusing on distinguishing between patients with Alzheimer’s Disease (AD) and cognitively normal controls.

• The Age of AI and Overconfidence: The effects of domain expertise on AI trust

  This study observes the effects of domain expertise on AI trust.

• The Dark Side of Algorithmic Adaptivity: How Model Recommendations Shape Consumer Vulnerability

  This project advances on ethical AI by integrating consumer behavior and marketing to examine how consumers experience and interpret adaptive AI systems over time.

• Understanding How Neurodiverse Learners Use AI-Assisted Programming Tools with Eye tracking Studies

  The goal of this project are to study opportunities and challenges that existing AI-assisted programming tools pose to neurodiverse students.