Obvia awards 13 new scholarships to support emerging researchers, from college to doctoral level

Research Project: Promoting Physical Activity Through AI Interaction in CEGEPs: An Interdisciplinary Approach to Student Health

This research project aims to explore the use of artificial intelligence to accurately and effectively assess students’ performance during specific physical exercises, such as squats and push-ups. The goal is to develop an AI-based system capable of analyzing the form and execution of movements and ensuring their accuracy.

Research Project: When AI Saves Lives: Emerging Uses of Generative AI in Medical Emergency Situations

This project explores the use of generative artificial intelligence, specifically ChatGPT, by non-clinicians in medical emergencies. Through a literature review, analysis of social media comments (Reddit, TikTok, X), and a comparative analysis of certain AI models (ChatGPT, Gemini, Claude, etc.), it aims to assess the applicability, consistency, accuracy, and ethical implications of AI-generated responses in critical situations such as chest pain, allergic reactions, or overdoses.

Interviews with healthcare professionals will enrich the analysis. The project also addresses issues of digital governance, misinformation, informed consent, and data protection. It aligns with OBVIA’s mission by examining the social implications of AI and contributing to the responsible, human-centered development of these technologies in the field of public health.

Research Project: Toward a Better Understanding of the Integration of AI Tools and Their Impact on Human Work in an Organizational Context

This project examines the emergence of generative artificial intelligence (AI) and the regulations governing its use, which have sparked significant interest in how AI can support work, particularly in performing office tasks (e.g., data analysis, report writing, process automation, etc.) Among the most widely used AI tools are conversational agents (chatbots), designed to save users time and cognitive effort. Although it is recognized that these AI tools can serve as complementary tools to work performed by humans, how this collaborative work is actually carried out, as well as how individuals actually interact with AI tools, remains unknown. This project, consisting of a scoping study, an experimental study, and a qualitative study, aims to develop a more comprehensive understanding of the integration and use of AI tools in an office work setting.

Research Project: Administrative Courts and Artificial Intelligence: Implementing a Decision-Support System at the TAL

The Administrative Housing Tribunal (TAL) is a court that adjudicates disputes in the area of residential leasing, with average resolution times ranging from several months to several years. Consequently, it is essential to identify solutions to support this court. This project aims to explore the possibility of improving the TAL’s efficiency using an AI-based decision-support tool that relies on judgment capture—a method for extracting an individual’s decision-making patterns and generating a predictive model from them. Administrative housing law is well-suited to this modeling strategy due to its formal procedural rules and the rule of the preponderance of the evidence. Decision-making models will be generated from case law and presented to TAL judges to obtain their feedback on the models’ potential to support decision-making.

Research Project: Acceptability of Virtual Agents in Healthcare Settings: The Role of the Uncanny Valley Phenomenon

This project aims to conduct a literature review—initially narrative in nature, with the possibility of evolving into a systematic review—to answer the following questions: 1) To what extent is the use of a virtual agent accepted when it fills a role typically reserved for a human in a healthcare setting, and 2) what factors influence this acceptability (e.g., visual realism, fluidity of movement, response time, attractiveness, etc.)? This review will then guide the design of an experimental task involving virtual agents, in order to empirically study the influence of the uncanny valley phenomenon on their acceptability.

Research Project: “Acting” Against Deception: A Psychodramatic Approach to Scam Awareness

This research addresses the growing threat of scams by developing an innovative framework that combines psychodrama, generative AI, and gamification. Unlike traditional awareness-raising methods, which often fail under emotional pressure, this approach immerses participants in realistic scam simulations, where they take turns playing the roles of both victim and scammer.

Through dynamic AI-generated scenarios and guided reflection, participants develop their emotional resilience and critical thinking skills. The system adapts to users’ behavior, rewarding effective decision-making and fostering lasting resistance to scams.

Designed to address the psychological vulnerabilities exploited by fraudsters, this framework aims to reduce actual victimization and promote digital well-being. It represents a novel and practical application of AI for cybersecurity education, in line with the principles of responsible innovation that emphasize human autonomy, safety, and mental health.

Research Project: From Governance to Public Trust: Exploring Ways for an External Party to Verify an AI System’s Compliance with Regulations and Ethical Standards

This project focuses on creating and evaluating methods that enable an external party to verify the compliance of proprietary AI systems with ethical expectations and regulations, without requiring the public disclosure of all their components. In light of the growing opacity of AI models and the limitations of open source, the project proposes to evaluate the reliability, cost, accessibility, and feasibility of existing methods for verifying compliance with ethical standards in proprietary AI systems, and then to adapt these methods or design new ones to address the shortcomings of existing approaches. This research project takes an interdisciplinary approach (AI, cryptography, law, and ethics). The goal is to strengthen corporate accountability, improve public trust, and make AI governance more practical and flexible.

Research Project: Programmed Bias and Artificial Intelligence: The Representation of Women in AI Systems

This project on algorithmic bias and the representation of women in the digital realm aims to critique current models for the development of artificial intelligence systems.

This research uses the philosophy of feminist Mary Wollstonecraft as a backdrop to highlight current issues regarding the representation of women in algorithm training databases as well as in the design of voice assistants. A dialogue between the 18th-century philosopher’s arguments and the current oppression of women in the digital realm demonstrates how artificial intelligence (AI) serves as a vehicle for amplifying inequalities already present in the physical world.

The research presents avenues for reflection on the consequences of these representations in the digital sphere. It analyzes the risks of harassment, violence, and radicalization targeting women.

Research Project: How Can We Contribute to the Beneficial Use of Algorithmic Tools in Predictive Policing Policies and Practices?

Predictive algorithmic technologies, including artificial intelligence, are rapidly being integrated into policing practices, often without prior study of their concrete effects. While their adoption promises efficiency and prevention, it also raises critical issues: bias, opacity, and the erosion of human accountability. This observation leads to a crucial question: What is the impact of using algorithmic tools in the field of predictive policing in Quebec? This project aims to explore this issue by analyzing how these tools are adopted and used in the field, as well as the discrepancies between the strategic objectives of their implementation and their actual use by police officers. The methodology is based on ethnographic research within the SPVM, as well as semi-structured interviews with police officers, decision-makers, and key informants. Ultimately, this research will provide an empirical foundation to guide the development of responsible public policies tailored to on-the-ground realities.

Research Project: Developing Responsible Artificial Intelligence to Reduce HIV-Related Stigma and Promote Sustainable Psychosocial Care

This project aims to strengthen mental health support for people living with HIV (PLHIV) by developing responsible artificial intelligence integrated into MARVIN, a chatbot designed to support these individuals between medical appointments. The goal is to enable MARVIN to automatically detect signs of emotional distress or stigmatizing language in users’ messages and then respond with empathy, using inclusive and respectful language. The project is based on a participatory approach that directly involves patients, clinicians, and engineers at every stage of design. It is rooted in a vision of sustainable health and digital ethics, seeking to make technologies more humane, accessible, and equitable. By fostering a caring and tailored dialogue, this innovation will improve the quality of care, reduce stigma, and restore patients’ trust in the healthcare system.

Research Project: AI-Driven Personalized Exercise Planning for People Living with HIV: An Intervention to Improve Physical Activity and Health Outcomes through the MARVIN Chatbot

As people living with HIV (PLWHIV) age, it is essential to support their physical and mental health. Despite the recognized benefits of exercise, only half of PLWHIV meet Canadian physical activity guidelines. Barriers to exercise include uncertainty about how to start or maintain a routine, as well as limited access to trained professionals, who are often expensive.

To address these challenges, I am working on developing an exercise module within MARVIN, an AI-based chatbot developed by our team (B. Lebouché, principal investigator). Using a decision tree built on the principles of exercise science and prompt engineering, the chatbot will generate personalized 8-week training plans tailored to patients’ preferences, risk factors, and activity levels.

The effectiveness of this module will be evaluated through tests with patient partners and pilot studies. The outcomes examined will include changes in exercise-related behaviors, mood, energy levels, and motivation. MARVIN represents a promising and scalable approach to helping people living with HIV—as well as others living with chronic conditions—adopt a regular exercise routine, without increasing the burden on healthcare professionals.

Research Project: Integration of Generative Artificial Intelligence Technologies: What Impact Do They Have on the Leadership of Public Managers?

This research project explores the impact of generative artificial intelligence (GAI) on the leadership practices of public managers in Quebec. As GAI transforms decision-making practices in public administration, it raises unprecedented tensions between automation, human accountability, and managerial legitimacy. Despite a growing regulatory framework, the concrete effects of GAI on managers’ roles, competencies, and practices remain poorly documented. Using a mixed-methods approach—including a systematic review, case studies, and co-design—this project aims to generate rigorous empirical knowledge on digital leadership while co-developing an operational analytical tool. Rooted in a perspective of ethical governance and responsible innovation, it integrates insights from public administration, the sociology of work, and the ethics of AI. Ultimately, it seeks to support a reflective, equitable, and context-sensitive managerial transformation, helping to strengthen the capacity of public institutions to act in the algorithmic era.

Research Project: Strengthening Digital Equity in Health Care: Co-Development of Support Interventions for Migrants with Precarious Status in Quebec

Access to health care in Quebec increasingly relies on digital tools. However, for migrants with precarious status (asylum seekers, temporary workers, undocumented individuals), these technologies can pose significant barriers. The project aims to identify the obstacles encountered in using digital health services and to co-develop, with these populations, tailored tools (visual guides, short videos, workshops). A literature review will be followed by interviews and focus groups with migrants, community workers, and healthcare professionals. A pilot phase will test the tools developed and evaluate their clarity, usefulness, and reach. Based on a participatory and interdisciplinary approach (public health, social sciences, technology), the project mobilizes a network of community partners. It will help document digital health disparities and propose concrete, scalable, and inclusive solutions to improve equitable access to care.