AI's Social Forcefield: Reshaping Distributed Cognition in Human-AI Teams

TL;DR

This paper reframes AI from a passive tool to an active participant that reshapes distributed cognition in human-AI teams. Through two randomized studies, it demonstrates that exposure to AI-generated language causally alters lexical choices, attention focus, mental models, and group identity, with effects persisting into human-only interactions. The authors introduce a unified framework of distributed cognitive alignment and depict AI as a social forcefield that operates across micro-, meso-, and macro-level coordination. While AI can improve alignment and coordination, it can also erode epistemic diversity and homogenize expression, highlighting the need for transparent, controllable design that accounts for group dynamics. The work calls for a paradigm shift in AI design and evaluation, prioritizing social impact and long-term cognitive ecology of teams.

Abstract

AI is not only a neutral tool in team settings; it actively reshapes the social and cognitive fabric of collaboration. We advance a unified framework of alignment in distributed cognition in human-AI teams -- a process through which linguistic, cognitive, and social coordination emerge as human and AI agents co-construct a shared representational space. Across two studies, we show that exposure to AI-generated language shapes not only how people speak, but also how they think, what they attend to, and how they relate to each other. Together, these findings reveal how AI participation reorganizes the distributed cognitive architecture of teams: AI systems function as implicit social forcefields. Our findings highlight the double-edged impact of AI: the same mechanisms that enable efficient collaboration can also erode epistemic diversity and undermine natural alignment processes. We argue for rethinking AI in teams as a socially influential actor and call for new design paradigms that foreground transparency, controllability, and group-level dynamics to foster responsible, productive human-AI collaboration.

Figures (8)

• Figure 1: AI exerts a "social forcefield," affecting the cognitive alignment in human groups through a multi-level model, from Micro-level coupling of linguistic entrainment, through meso-level alignment of attention and shared mental models, to macro-level alignment through relational and social integration.
• Figure 2: Three chests from the Cursed Treasure puzzle. A treasure chest is cursed with the "octopus curse" if the gems are separated but not cursed if the gems are touching.
• Figure 3: Illustration of lexical alignment on the AI assistant's language in one example team.
• Figure 4: AI affects the shared focus of attention. Teams talk more about the object mentioned by the AI in the segment immediately after its interjection than at other times. Counts include all mentions of tracked objects, regardless of which terms are used.
• Figure 5: Pronoun use indicates strong team cohesion in the Robotic-Unhelpful condition (based on successful exclusion of AI), yet disrupted team cohesion in the Human-Unhelpful condition (based on unsuccessful exclusion of AI). Bars show count of each pronoun across all teams in each treatment.