Submissions

What can narrative agents do for creativity and personal well being? I come to this space not as someone versed in agentic AI but from the narrative generation side of AI. We use AI in my lab to study human storytelling, which necessitates a formal understanding of the nature and purpose of narrative communication. The question I bring to the workshop is how might we design AI storytelling agents who are able to work with human users to further their creativity and support their mental health? Fears of AI weakening or usurping our creativity are rampant. At the same time, we know "how we tell our own story" is tightly coupled with our sense of well-being. I am seeking collaborations that help move us toward this healthy, human-centred space of storytelling AI.

I am excited to participate in the ACM COMPASS workshop to explore how social theory can meaningfully inform the design and evaluation of agentic AI systems situated in real-world organizational contexts. My work bridges AI development and sociotechnical analysis, and I look forward to contributing insights on integrating social perspectives into responsible agentic system design.

I have a work-in-progress position paper, but it is not ready to be shared. Here's an abstract that gives an overview of my background and interests, though. I hope to discuss some of these topics and potentially look for future collaborators, though. 

Working title: Parallel Practices for Sound-Interaction: Using Urban Soundscape Tools to Consider the Future Role of AI

Sound remains an underrepresented dimension in urban planning, often treated as a regulatory afterthought rather than a meaningful aspect of urban experience. This position paper discusses several parallels found in during the development of new tools for urban sound planning, which would be relevant for any attempt of introducing any form of AIs into the process.

Urban sound planning remains a global challenge and under-recognized sustainability goal, as current practice lags for behind research. In short, sound is often ignored, taken for granted, or deemed to complex and passed off to experts. These experts, similarly, do not have enough tools, knowledge, or regulatory support to go above and beyond a sound-level centric status quo that considers sound as noise to be mitigated. Yet, sound is much more than noise! Current practices rarely consider using sound as a resource and have very few resources to plan for sound in early stages. This is further complicated as planning for an auditory experience cannot directly be measured by sensors. Likewise, this is further complicated when considering culturally distinct locales, for shared public spaces. Auditory – or more generally urban experiences – cannot be separated from place, and is indeed, an essential element of one’s sense of place.

Such complex issues are at the heart of third wave HCI as they seek to understand root causes and provide solutions to help people consider societal issues and support working through them. Several experiences are shared from the R&D of a new immersive soundscape planning tool, City Ditty, developed in response to these issues. It is with hope that sharing these will help spread sound-awareness to highlight global issues around sound, introduce several challenges of working with sound to help address, and help identify points for collaboration with AI.

More info: https://www.sounds-in-the-city.org/en/ 

I am currently working on a project with the ESRC sociological futures hub on the future of agents within family life. Our preliminary research shows tensions that arise between parents and children in relation to technologies in the home, and in particular smart home technology such as Alexa. We are looking to probe families’ views on the future of agents within family life using research through design of multi-agent, multiple family member users. Matt’s keynote at CUI was very thought provoking, and I am hoping this workshop on social context of agents will help inform future research directions,

I’m interested in the use of agentic systems in a number of settings including the home, and the workplace and in music creation and in end user engagement in designing these systems. 

My research develops stakeholder-driven AI tools for watershed management that must navigate the complex social dynamics between community members, planners, and technical experts who collectively steward these systems. As I work toward developing custom MCP frameworks for enhanced tool integration, I'm particularly interested in how agentic AI can recognize and respond to the situated knowledge of different stakeholder groups within watershed networks. This work directly addresses the workshop's emphasis on socially-situated AI by examining how environmental AI systems must account for the social relationships, power dynamics, and local expertise that shape community water resource management.

Addressing the 'Invisibility' Challenge: Designing and Situating Agentic AI within Social Context for Older Adults and Children

Rezvan Boostani, PhD Candidate, Faculty of Information, University of Toronto, rezvan.boostani@mail.utoronto.ca

My interest in the ACM COMPASS workshop on “Social Agentics” stems from a critical challenge in designing AI for older adults and children: their (their context) "invisibility" to designers and technologists [3, 7, 8]. While agentic AI promises enhanced independence and access to services for these populations [1, 3, 4, 5, 11], current understanding of their unique contexts and needs is limited [2, 6]. This gap is due to delegated online identities and facilitated access, where system logs reflect intermediaries' and networks of family and friends' activities rather than the actual users'. For instance, older adults often rely on their network to perform online tasks including but not limited to finance management, healthcare information access, shopping, and socialization. This creates a false sense of user understanding and success metrics for designers.

The delegated nature of technology use creates inaccurate assumptions of older adults and children's needs and context for designers and technologists [4, 9, 10]. This leads to the creation of agentic systems that are not adequately tailored to their unique needs and context. It also has implications for data collection and analysis, as the user's behavior is not captured. 

I aim to explore how we can gather authentic usage data and develop design strategies that cater to needs of these groups, ensuring responsible and equitable AI development. I believe this workshop offers a vital platform to collaboratively discuss and imagine agentic AI that addresses these implications for designing age-inclusive agentic AI systems. 

To develop responsible approaches for social agentics, we must address this "invisibility." Key questions for design:

• What design strategies can account for delegated interactions while still serving the primary user?

• In what ways do agentic systems cater to the needs and context of older adults, rather than further marginalizing them? 

References

1. Agentic AI: Developing the benefits for classroom learning – part I (no date) NSTA. Available at: https://www.nsta.org/blog/agentic-ai-developing-benefits-classroom-learning-part-i#:~:text=Summarization%3A%20Condensing%20key%20information%20from,Adapting%20to%20the%20student’s%20understanding. (Accessed: July 2025). 

2. Bran Knowles, Vicki L Hanson, Yvonne Rogers, Anne Marie Piper, Jenny Waycott, Nigel Davies, Aloha Hufana Ambe, Robin N Brewer, Debaleena Chattopadhyay, Marianne Dee, et al. 2021. The harm in conflating aging with accessibility. Commun. ACM 64, 7 (2021), 66–71.

3. Choi, J., Choi, S., Song, K., Baek, J., Kim, H., Choi, M., ... & Shin, J. (2023). Everyday digital literacy questionnaire for older adults: instrument development and validation study. Journal of medical Internet research, 25, e51616.

4. El Sherif, R., Pluye, P., & Ibekwe, F. (2022). Contexts and outcomes of proxy online health information seeking: mixed studies review with framework synthesis. Journal of Medical Internet Research, 24(6), e34345. 

5. Gosmar, D., & Dahl, D. A. (2025). Hallucination mitigation using agentic ai natural language-based frameworks. arXiv preprint arXiv:2501.13946.

6. Kelly Mack, Emma McDonnell, Dhruv Jain, Lucy Lu Wang, Jon E. Froehlich, and Leah Findlater. 2021. What do we mean by “accessibility research”? A literature survey of accessibility papers in CHI and ASSETS from 1994 to 2019. In Proceedings of the 2021 CHI Conference on Human Factors in Computing Systems. 1–18.

7. Shade, M., Yan, C., Jones, V. K., & Boron, J. (2025). Evaluating older adults’ engagement and usability with AI-driven interventions: Randomized pilot study. JMIR Formative Research, 9, e64763.

8. Sehgal, G. (2025) AI & Agentic AI in education: Shaping the future of learning, Medium. Available at: https://medium.com/accredian/ai-agentic-ai-in-education-shaping-the-future-of-learning-1e46ce9be0c1 (Accessed: 12 July 2025). 

9. Thalhammer, D. V. (2024). Family support for older adults’ media use: Implications for COVID-19 research. Journal of Intergenerational Relationships, 1-18. 

10. Webster, G., & Ryan, F. V. (2023). Social Media by Proxy: how older adults work within their ‘social networks’ to engage with social media. Information Research: an international electronic journal, 28(1), 50-77. 

11. Wong, A. K. C., Lee, J. H. T., Zhao, Y., Lu, Q., Yang, S., & Hui, V. C. C. (2025). Exploring older adults’ perspectives and acceptance of ai-driven health technologies: qualitative study. JMIR aging, 8, e66778.