More Than 1v1: Human-AI Alignment in Early Developmental Communities with Multimodal LLMs

TL;DR

This paper explores how multimodal large language models can support early developmental community interactions, emphasizing layered community alignment involving families, professionals, and AI systems.

Contribution

It introduces the concept of layered community alignment, proposing a community-governed approach to AI support in developmental contexts.

Findings

MLLM outputs transition from expert analysis to parent feedback

Layered community alignment involves grounding, translation, and adaptation

Alignment should be community-governed, not just individual optimization

Abstract

In early developmental contexts, particularly in parent-child interaction analysis, alignment involves families and professionals such as speech-language pathologists (SLPs) who interpret children's everyday interactions from different roles. When multimodal large language models (MLLMs) are introduced to support this process, alignment becomes a question of how authority, responsibility, and emotional risk are distributed across stakeholders. Through a three-part study with five families and three SLPs, we trace how MLLM-generated outputs move from expert-facing analysis to parent-facing feedback. We propose layered community alignment: grounding representations in expert-aligned structures, mediating translation through professional guardrails, and enabling family-level adaptation within those boundaries. We argue that alignment in developmental settings should be treated as a community-governed process rather than an individual optimisation problem.