Mimetic Alignment with ASPECT: Evaluation of AI-inferred Personal Profiles

2026-03-27

AI Summary

This work addresses the limitations of existing AI agents in modeling individual communication styles—often constrained by costly fine-tuning, superficial personas, or a focus on preferences while neglecting expressive nuances. The authors propose ASPECT, a novel framework that, for the first time, leverages large language models to automatically infer psychologically validated communication trait dimensions from workplace behavioral data without requiring fine-tuning, thereby generating interpretable and auditable personal communication profiles. Integrating psychological scales, prompt engineering, and behavioral analysis, ASPECT establishes an automated pipeline for social-psychological assessment. In a case study with 20 participants, ASPECT-generated profiles showed moderate alignment with self-reports, produced responses outperforming both generic and self-reported baselines, and featured an evidence-linking mechanism that effectively enabled users to calibrate AI representations, enhancing authenticity and controllability in expressive output.

Abstract

AI agents that communicate on behalf of individuals need to capture how each person actually communicates, yet current approaches either require costly per-person fine-tuning, produce generic outputs from shallow persona descriptions, or optimize preferences without modeling communication style. We present ASPECT (Automated Social Psychometric Evaluation of Communication Traits), a pipeline that directs LLMs to assess constructs from a validated communication scale against behavioral evidence from workplace data, without per-person training. In a case study with 20 participants (1,840 paired item ratings, 600 scenario evaluations), ASPECT-generated profiles achieved moderate alignment with self-assessments, and ASPECT-generated responses were preferred over generic and self-report baselines on aggregate, with substantial variation across individuals and scenarios. During the profile review phase, linked evidence helped participants identify mischaracterizations, recalibrate their own self-ratings, and negotiate context-appropriate representations. We discuss implications for building inspectable, individually scoped communication profiles that let individuals control how agents represent them at work.

Problem

• mimetic alignment
• communication style
• AI-inferred personal profiles
• persona modeling
• individual representation

Innovation

• mimetic alignment
• communication style modeling
• LLM-based psychometric evaluation
• inspectable AI profiles
• personalized AI agents