← CHI 2026 map

CHI '26 · Honorable mention · full-paper review · confidence medium-high

Futuring Social Assemblages: How Enmeshing AIs into Social Life Challenges the Individual and the Interpersonal

Lingqing Wang , Yingting Gao , Chidimma Lois Anyi , Ashok Goel

DOI PDF Program page Page JSON

This is a thoughtful CHI position paper grounded in a small qualitative design study. Its strongest contribution is not a new system but a reframing: social AI should be judged by interpersonal and secondary-user consequences, not only by the primary user’s experience.

Axes Lens

Rare contribution shape, typical evidence profile. The point here is not a score. It is to show what kind of claim the paper makes, and whether the evidence pattern is unusual or baseline in this 268 -review set.

Review Summary

This paper’s value lies in how it reorients the design conversation around social AI. Rather than treating AI as a tool that simply helps an individual user accomplish a task or feel better, it argues that once AI is woven into social life, the relevant unit of analysis expands to include interpersonal relations, secondary users, and longer-term identity and agency effects. That is a meaningful departure from common-sense product framing, and the abstract makes the argumentative stance explicit: social AI can exacerbate the problems it is meant to mitigate, introduce privacy harms for bystanders or other participants, and threaten the primary user’s sense of self-efficacy and identity. The evidence base is a three-phase design study with 24 participants, which is appropriate for surfacing tensions and generating a conceptual reframing, but it is not strong evidence for broad causal claims about deployed systems. In other words, the paper is best read as a normative and interpretive contribution: it synthesizes participant responses to speculative scenarios into a field-level argument for a relational design perspective. The novelty is therefore primarily conceptual/framework-level rather than technical or empirical in the sense of a new algorithm or benchmark. The limitations matter: the study is qualitative, scenario-based, and bounded by a homogeneous participant pool and narrow social context, so the findings should not be overgeneralized to all users or all social AI settings. Still, within CHI, this is a credible and timely contribution because it names a real gap in user-centered design discourse and offers a sharper vocabulary for thinking about harms that extend beyond the primary user. The paper’s strongest claim is not that it proves social AI is harmful, but that current design paradigms can miss important relational costs unless they explicitly account for them. The discussion and limitations sections reinforce that reading by explicitly noting the narrow participant pool, the educational/professional context, and the speculative nature of the design fictions. Taken together, the paper is best understood as a field argument: it does not deliver a deployable system, but it does provide a persuasive conceptual lens for future social-AI design and evaluation.

What Changed

Canon before

CHI work on social AI and user-centered design typically emphasizes immediate user experience, individual benefit, and direct interaction outcomes; this paper shifts attention to interpersonal ethics, secondary users, and longer-term social consequences.

Departure from common sense

The paper’s core move is counterintuitive: instead of assuming social AI improves social life, it argues that social AI can worsen the very interpersonal problems it is intended to solve and can also create harms for people beyond the primary user.

Actual novelty

Its main contribution is a relational design perspective that treats social AI as embedded in social assemblages, foregrounding secondary users, interpersonal ethics, and long-term personal consequences as first-class design concerns rather than peripheral side effects.

Evidence

The paper reports a three-phase design study with 24 participants and uses speculative design fictions/storyboards plus interviews and thematic analysis to surface tensions around interpersonal harm, privacy for secondary users, and threats to agency and identity. The evidence supports a qualitative, interpretive argument rather than deployment-level causal claims.

Limits

Method limits

The study is qualitative and speculative rather than a controlled or field deployment evaluation, so it supports participant reactions and interpretive themes more than causal claims about actual system effects.

Deployment limits

Because the evidence comes from speculative design materials and interviews, the findings do not directly establish how deployed social AI systems will behave in real-world settings or across diverse populations.

Boundary conditions

The claims are bounded by the study’s participant pool and scenario-based elicitation; they are most applicable to social AI contexts where interactions involve multiple people and where secondary-user effects and identity/agency concerns are salient.

Manual check: false

Position in field

The paper positions itself against a narrow user-centered paradigm and toward a relational, socially situated design agenda for AI systems that mediate human relationships. It contributes a field-level critique of optimizing only for the primary user, arguing that social AI should be evaluated through interpersonal ethics, secondary-user impacts, and longer-term consequences that extend beyond immediate interaction quality.

Abstract

Recent advances in AI are integrating AI into the fabric of human social life, creating transformative, co-shaping relationships between humans and AI. This trend makes it urgent to investigate how these systems, in turn, shape their users. We conducted a three-phase design study with 24 participants to explore this dynamic. Our findings reveal critical tensions: (1) social AI often exacerbates the very interpersonal problems it is designed to mitigate; (2) it introduces nuanced privacy harms for secondary users inadvertently involved in AI-mediated social interactions; and (3) it can threaten the primary user's personal agency and identity. We argue these tensions expose a problematic tendency in the user-centered paradigm, which often prioritizes immediate user experience at the expense of core human values like interpersonal ethics and self-efficacy. We call for a paradigm shift toward a more provocative and relational design perspective that foregrounds long-term social and personal consequences.