Are Conversational AI Agents the Way Out? Co-Designing Reader-Oriented News Experiences with Immigrants and Journalists

Paper Title

Are Conversational AI Agents the Way Out? Co-Designing Reader-Oriented News Experiences with Immigrants and Journalists

Publication Info

• Topic area: Enhancing news experiences for immigrant readers through co-design with journalists and conversational AI.
• Keywords: Conversational AI, news consumption, immigrant readers, co-design, human-centered computing, journalism, reader-oriented design, value alignment, human-AI collaboration, marginalized readership.

Background and Problem

• Problem / challenge: Immigrant readers face significant barriers in engaging with mainstream news due to language fluency, cultural unfamiliarity, and mismatches with local societal norms. Existing solutions often treat immigrant readers in isolation and fail to integrate journalists’ perspectives, leaving a gap in creating inclusive, reader-oriented news experiences.
• Significance: Addressing these challenges is crucial for improving immigrants' access to reliable information, fostering societal integration, and mitigating risks associated with misinformation and echo chambers.
• Motivation and related work: Prior work has explored tools like automated translation, accessible text frameworks, and comprehension support from LLMs. However, these efforts often focus solely on technical solutions for immigrants without considering journalists' roles or fostering collaboration between stakeholders.

Solution

• Proposed approach: Co-designing news technologies with immigrant readers and journalists to create conversational AI agents that enhance news experiences while addressing value misalignments.
• Novelty:
  1. Identification of value (mis)alignments between immigrant readers and journalists in news topic selection, presentation, and processing.
  2. Development of four design metaphors for conversational AI agents: Data Decoder, Connection Informer, Empathetic Friend, and Trajectory Witness.
  3. Introduction of human-AI workflows that balance responsiveness to immigrant needs with journalistic accountability.
  4. Framework for collaborative, cross-stakeholder design of reader-oriented news technologies.
• Procedure and key techniques:
  • Conducted co-design activities with 11 immigrant readers and 7 journalists in the U.S. capital region.
  • Organized two activity sets: "News On My Radar" (exploring value alignments) and "News Without Barriers" (envisioning AI-assisted solutions).
  • Used thematic analysis to synthesize findings and derive design metaphors.

Results

• Concrete findings:
  • Immigrant readers prioritize actionable, context-rich, and empathetic news content tailored to their unique social and legal circumstances.
  • Journalists emphasize professional gatekeeping, credibility, and cautious presentation to maintain accountability.
  • Four design metaphors for conversational AI agents:
    1. Data Decoder: Explains numerical data and its context in news.
    2. Connection Informer: Highlights actionable implications of news for immigrants.
    3. Empathetic Friend: Reduces emotional strain by balancing negative news with positive content.
    4. Trajectory Witness: Analyzes and reflects on readers’ news consumption patterns over time.
• Advantage over baselines:
  • Bridges the gap between immigrant readers and journalists by fostering mutual understanding and collaboration.
  • Proposes AI solutions that go beyond isolated technical fixes to address systemic issues in news consumption.
• Experiments / evaluation:
  • Co-design sessions included diary studies, small-group discussions, and large-group ideation activities.
  • Data analysis revealed shared and divergent values, leading to actionable design insights.
• Limitations and future work:
  • Findings are limited to the local media landscape and participants' immediate contexts.
  • Future work should explore long-term dynamics of news practices and broader applicability of the proposed metaphors.

Summary

This study investigates how conversational AI agents can enhance news experiences for immigrant readers by addressing value misalignments with journalists. Through co-design activities with 11 immigrants and 7 journalists, the authors identified four design metaphors—Data Decoder, Connection Informer, Empathetic Friend, and Trajectory Witness—that outline distinct roles for AI in supporting news consumption. The proposed human-AI workflows balance immigrants' needs for actionable, empathetic content with journalists' commitment to accountability. These findings provide a foundation for creating inclusive, reader-oriented news technologies that integrate diverse stakeholder perspectives.