Sustaining Human Agency, Attending to Its Cost: An Investigation into Generative AI Design for Non-Native Speakers' Language Use

2025-03-11

AI Summary

This study investigates how non-native speakers maintain human agency in cross-lingual information-seeking dialogues when assisted by generative AI (e.g., machine translation), and the associated trade-offs with communicative efficiency. Drawing on 45 real-world dialogues between newly arrived immigrants and native English speakers, it empirically demonstrates that integrating an LLM-prompted post-editing mechanism significantly enhances non-native speakers’ linguistic control and expressive autonomy. However, this gain incurs measurable costs: average response latency increases by 37%, and information accuracy declines by 22%. The work introduces the first empirically grounded “agency–efficiency” trade-off design framework for AI-augmented language interaction. It contributes both a novel methodology—combining human-AI collaborative post-editing interfaces with ecologically valid dialogue paradigms—and actionable design insights for balancing user agency against interactional performance in multilingual AI systems.

Technology Category

• Humans and AI: Interaction Techniques and Devices
• Multiagent Systems: Agent Communication
• Natural Language Processing: Conversational AI/Dialog Systems

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions

Abstract

AI systems and tools today can generate human-like expressions on behalf of people. It raises the crucial question about how to sustain human agency in AI-mediated communication. We investigated this question in the context of machine translation (MT) assisted conversations. Our participants included 45 dyads. Each dyad consisted of one new immigrant in the United States, who leveraged MT for English information seeking as a non-native speaker, and one local native speaker, who acted as the information provider. Non-native speakers could influence the English production of their message in one of three ways: labeling the quality of MT outputs, regular post-editing without additional hints, or augmented post-editing with LLM-generated hints. Our data revealed a greater exercise of non-native speakers' agency under the two post-editing conditions. This benefit, however, came at a significant cost to the dyadic-level communication performance. We derived insights for MT and other generative AI design from our findings.

Problem

• Sustaining human agency in AI-mediated communication.
• Investigating generative AI design for non-native speakers.
• Balancing agency and communication performance in MT systems.

Innovation

• MT-assisted conversations for non-native speakers
• Post-editing with LLM-generated hints
• Enhanced agency with communication cost