We asked the two students who were selected for FY2025 Integrated Research Project about various aspects of the project. Please take a look!
【Researcher Name】
Momoka Matsuse (Graduate School of Law, Department of Law and Politics)
SHAH MANAN VINOD (Graduate School of Information Science and Electrical Engineering, Department of Information Science and Technology)
【Integrated Research Title】
International Comparison of Ethics and Legal Issues in AI Diagnosis
【Research Abstract】
This proposed research aims to compare AI diagnosis regulations across multiple countries to propose an ideal legal framework for AI in healthcare. As AI-driven diagnostic tools become increasingly prevalent, regulatory disparities pose challenges to cross-border implementation and patient safety. The research will employ a multi-method approach, including data collection of existing regulations, Natural Language Processing (NLP) to extract key themes from policy documents, and a case study comparison to analyze legal disputes related to AI diagnosis.
The study will focus on issues such as liability, transparency, data privacy, and ethical considerations in AI-based medical decisions. By identifying commonalities and differences in regulatory approaches, the research aims to propose a comprehensive and harmonized set of policy recommendations. Expected outcomes include a database for AI medical law comparison and policy proposals for standardizing AI regulations globally. Such consolidation of the knowledge base aims to facilitate safer and more efficient adoption of AI in healthcare while addressing ethical and legal challenges. The findings could serve as a foundation for international cooperation on AI governance in the medical field, promoting access to AI-driven healthcare solutions.
Q1. What prompted you to start this joint research?
Our collaboration began in the Fusion of Intelligence Course A, where we shared an interest in safer, fairer AI for healthcare and saw how our skills complemented each other, law/ethics (Matsuse) and AI/engineering (Manan). We identified a gap: fragmented, fast-changing rules for AI across countries. The idea was to compare Japan, the EU, and the US, extract common themes (liability, transparency, privacy), and propose a harmonized legal question answering system that researchers and clinicians could use. The project concept crystallized into two tangible outputs: a comparative database plus policy recommendations.
Q2. How did you feel about being selected for the Integrated Research Project?
We felt validated, the selection signaled that bridging AI engineering and legal information is valuable for assisting policymakers and even for common people to understand policy information, adding a layer providing ease of access. It also created a healthy pressure to deliver the two core outputs we proposed, a comparative database and policy question answering. The grant structure gave us confidence to aim for conferences, not just internal results. In short, gratitude + responsibility: we now have resources to execute what we envisioned.
Q3. What is the current progress of your joint research?
Technically, we created the environment, database, logic and multilingual chat application. A translation layer enables Japanese sources to be searched alongside English ones. On the legal side, a curated literature list (24+ sources incl. papers, books, guidelines) is organized; new items continue to be added and verified. We’ve begun showing comparative analysis alongside the steps taken by the system for extracting answers from multiple regions. Validation and test planning are in progress.
Q4. What new “insights” have you gained while conducting joint research?
We realized that policy documents pose unique challenges for analysis: each page often carries its own references, complicating data cleaning while preserving meaning. Another insight was how terminology varies across jurisdictions, terms used in English may not align directly with Japanese legal concepts, requiring contextual interpretation. These experiences showed us that effective analysis must account for both structural complexity and linguistic nuance. Importantly, we learned that combining technical methods with legal expertise is essential to capture these subtleties and produce accurate, interdisciplinary insights.
Q5. What are some of the difficulties of joint research that you did not think of when you were planning (applying for) this project?
Currently, we are focused on collecting data from the US, EU, and Japan. While we expected journals and policy documents to be organized by country, many papers combine information across regions. This overlap has made it harder to split and assign data cleanly between regions, requiring additional effort to disentangle region-specific content before analysis.
Q6. What is the reaction of your academic advisors and the members around you in the lab?
The professors have responded with both interest and recognition of the challenges. On the engineering side, there is enthusiasm about how data-driven analysis can reveal meaningful insights for regulation and practice. On the legal side, the reaction has highlighted the difficulty of working through vast amounts of policy documents, progress often depends on carefully reading each text rather than accessing ready-made datasets and generative answers. Lab members also find the interdisciplinary approach refreshing, as it opens new angles of discussion.
Q7. How do you actually conduct this joint research? (online, get together in a lab, etc.)
We meet weekly in person to test some built features, identifying challenges and opportunities together. For documentation/data sharing, we make use of Slack, and emails. Importantly, we connect with our mentors monthly through online sessions to share progress and get feedback. This structure, weekly collaboration, daily online coordination, and monthly mentorship ensure steady progress while balancing our different expertise and schedules.
Q8. How did you spend your research budget? What could you do/what would you like to do if you had another research budget?
We plan to use about 10% of our budget on OpenAI services for NLP tasks, while the remaining funds are dedicated to conference participation or journal publishing. With additional budget, we would deploy our application on a cloud environment, making it globally accessible and enhancing its impact.
Q9. Do you think there is a good chance of presenting this joint research at a conference or writing a paper? What percentage do you think there is?
Yes, we believe there is a strong chance. Based on our current progress and our interdisciplinary approach, we estimate an 80% probability of presenting our work at a conference or publishing it in a journal.
Q10. What are your future prospects and aspirations for this joint research?
Our current research focuses on the U.S., EU, and Japan, but our aspiration is to expand globally. AI in healthcare is a worldwide challenge, and including more regions would capture diverse perspectives and unique concerns.
Q11. What are your dreams for the future and what kind of person do you both aspire to be?
(Matsuse) I aim to pursue a career where I can apply analytical thinking, make thoughtful decisions, and approach situations with empathy.
(Manan) I aspire to be an innovator who provides AI-enabled solutions for real-world problems, making AI more accessible and solution-driven for everyone.
Q12. Please leave a message for the program students who follow to take on the challenge of the Integrated Research Project and joint research.
Embrace the challenge of stepping outside your field and working across disciplines. Joint research may feel difficult, but it opens doors to perspectives and approaches impossible to gain alone. Through our project, we are learning to tackle complex problems in innovative ways, navigate the challenges of collaboration, and appreciate both the strengths and limitations of our disciplines. The process is demanding but ultimately rewarding, and we consider the experience an invaluable asset.
