Call for Participation
This one-day ACM COMPASS workshop invites a diverse group of participants to explore how and why to design agentic systems to be situated within specific social and organizational contexts. We are particularly interested in engaging developers of agentic systems as well as scholars engaged in work that might be relevant to better understanding how the ‘social’ matters in building and evaluating agentic AI.
Together, we will explore the value of social theory and perspectives to this work, and the potential of this move to address critical issues with AI. Our aims are to develop responsible approaches for the future development of social agentics, and to collaboratively imagine socially and organizationally specific agentic AI systems in a range of different areas of practice.
We invite you to participate! Share your work and perspectives, reflect on your practices, and engage with others in the community. Submit a position contribution that explores how we can better incorporate social knowledge, theory, and perspectives into the design, operation, and evaluation of agentic AI. Please see below in workshop themes for potential starting points. Please note that all attendees need to register for COMPASS: https://compass.acm.org/
Submission
Position contribution can be a short expression of interest (EoI) with one to two sentences in the submission form, a short video or slides on work-in-progress, or a position paper. Position papers should be 2-4 pages long (excluding references) in the ACM SIGCHI Master Article template. Please submit your EoI here.
If you have any questions, please contact Matt Ratto (matt.ratto@utoronto.ca) or Ali Sutani (ali.sutani@utoronto.ca).
Workshop Details
Date/time: Wednesday, July 23, 2025, 9:00 AM to 4:30 PM
Location: Schwartz Reisman Innovation Campus (108 College St.), Room W762
Participants can also join virtually through Teams.
Workshop Themes
The overall goal of the workshop is to begin to constitute a shared area of practice from which research collaborations and shared projects can be developed.
To guide the collective work of attendees, we start from two general questions:
How are AI agents currently interpolated within existing social and cultural systems such that they can operate successfully within them, and what are the best ways to track this?
How can a better understanding of agentic systems and socio-cultural knowledge and practice support new agentic architectures and design patterns?
See the full workshop paper for more details.