Connecting Research, Design, and Policy

AI Companion Relationships is a VIP team focused on identifying benefits and harms of AI companion relationships, with an additional focus on disrupting the harms and amplifying benefits through connecting research findings to public policies and platform designs. By AI companions, this refers to computer systems that mimic human conversation and include voice assistants (e.g., Alexa) and large language models (e.g., ChatGPT, Gemini, Claude).

Students will learn to contribute to and lead interdisciplinary teams. No experience necessary to join. Students will learn how to conduct literature reviews, create presentations, use statistical software, design research, and the appropriate use of large language models to assist with research.

Issues Involved or Addressed

Example research questions include:

• How do different features of AI companions influence loneliness?
• How can AI companions improve human-to-human social connection?
• What types of AI literacy is necessary to promote healthy AI companion relationships?
• How do you limit overreliance on AI companions for social connection?
• How do young people feel about banning minors from using AI companions?
• What are U.S. states doing to reduce harms from using AI companions?

Methods and Technologies

AWe use a range of research design, from content analysis to experiments to interviews, and use both human and automated methods to analyze data.

Members of our team will learn and use the following tools and technology:

• Qualtrics for survey design and data collection
• SONA Systems for participant recruitment
• Excel for data analysis
• R for data analysis
• Python for data analysis
• Large language models (e.g., Gemini, Claude)

Areas of Interest

• Information Science
• Games and Interactive Media
• Human-Computer Interaction
• Communication
• Psychology
• Sociology
• Public Health
• Human and Family Development
• Computer Science
• Law
• Political Science

Preferred Interests and Preparation

The team welcomes students from all academic backgrounds and levels of experience. We foster a collaborative, supportive, and exploratory learning environment. Group meetings are structured to encourage students to share different perspectives and contribute to idea generation.

Useful preparation and qualities may include:

• An interest in learning how to shape policy
• A motivation to design meaningful and healthy systems
• Openness for collaboration
• Eagerness to solve problems
• Respect for diversity in academic backgrounds, cultures, personal values, hobbies, and opinions within the team

Meeting Schedule & Location

• Meeting Time: Thursdays from 10:00-11:00am
• Location: Contact Dr. Campos-Castillo (camposca@msu.edu) for room location or Zoom link

Team Advisor

Dr. Celeste Campos-Castillo
Department of Media and Information, Michigan State University
camposca@msu.edu

Dr. Celeste Campos-Castillo is an Associate Professor in the Department of Media and Information at Michigan State University. She has co-authored with high school, undergraduate, and graduate students on research about how technologies shape health and health care outcomes. She regularly publishes in journals within the fields of data science, health policy, medicine, psychology, and public health. Her PhD is in Sociology from the University of Iowa and she completed a post-doctoral research fellowship at Dartmouth College.