Anthropomorphism and Trust in Human-Large Language Model interactions

2026-04-20 · Source: cs.AI updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Social Sciences & Behavioral Studies · Depth: Expert, extended

Summary

A study involving 115 participants and over 2,000 human-LLM interactions investigated how warmth, competence, and empathy influence anthropomorphism and trust in Large Language Models. Participants interacted with a Gemini 2.0-based chatbot, systematically varied in perceived warmth, competence, cognitive empathy, and affective empathy, across four topics: Biology, U.S. History, Lifestyle Management, and Relationship Advice. The research found that warmth and cognitive empathy significantly predicted perceptions of anthropomorphism, trust, similarity, relational closeness, frustration, and usefulness. Competence predicted all outcomes except anthropomorphism, while affective empathy primarily influenced relational measures but not epistemic outcomes like trust or usefulness. Subjective topics, such as relationship advice, amplified perceptions of human-likeness and relational connection more than objective topics, suggesting contextual factors play a role in user attribution of social traits to AI.

Key takeaway

For AI Product Managers developing conversational AI, understand that perceived warmth and cognitive empathy are crucial for fostering user trust and anthropomorphism across various interactions. While competence is vital for epistemic trust and usefulness, integrating appropriate levels of warmth and empathy, especially for subjective topics like relationship advice, can significantly enhance user engagement and relational closeness. Be mindful that excessive warmth without competence can reduce perceived human-likeness and increase frustration, necessitating a balanced approach to social trait design to avoid over-anthropomorphism and associated risks.

Key insights

Warmth, competence, and empathy are key dimensions shaping human perceptions of LLMs, influencing anthropomorphism and trust.

Principles

• Warmth and cognitive empathy drive broad positive perceptions of LLMs.
• Competence primarily builds trust and usefulness in LLMs.
• Subjective topics amplify human-likeness and relational connection with LLMs.

Method

Researchers systematically varied LLM chatbot responses for warmth, competence, and two types of empathy (cognitive, affective) using detailed prompts, then measured user perceptions across objective and subjective conversation topics.

In practice

• Prioritize competence for LLMs designed for factual tasks.
• Integrate warmth and cognitive empathy for relational AI applications.
• Tailor LLM social cues to topic subjectivity for enhanced user connection.

Topics

• Anthropomorphism
• Trust in AI
• Large Language Models
• Warmth and Competence
• Cognitive Empathy

Best for: AI Product Manager, AI Scientist, Research Scientist, AI Ethicist

Related on AIssential

• AI Can't Feel What Your Body Already Knows — Human interoception and physical social signaling provide unique advantages over AI in sensing and decision-making.
• The Decision to Verify: How Warmth and User Characteristics Shape Reliance on Conversational Agents for Information Search — User reliance on conversational AI persists despite fact-checking access, driven by prior trust and indirectly influenced by chatbot warmth.
• ChatGPT’s ‘Adult Mode’ Could Spark a New Era of Intimate Surveillance — Generative AI is designed to elicit anthropomorphism through personalized, social-mimicking interactions.
• When AI Says It Feels — LLMs can be trained to express human-like feelings and self-awareness via self-rewarded RL, with manageable performance trade-offs.
• Does Learning Require Feeling? Cameron Berg on the latest AI Consciousness & Welfare Research — AI systems exhibit growing evidence of introspection and functional emotions, suggesting potential for morally relevant subjective experiences.
• Imperfectly Cooperative Human-AI Interactions: Comparing the Impacts of Human and AI Attributes in Simulated and User Studies — AI transparency significantly impacts human trust and satisfaction in imperfectly cooperative interactions, often more than human personality.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.AI updates on arXiv.org.