What question did this study set out to answer?

This research aims to investigate the propensity to trust in large language models (LLMs) and understand how this trust is influenced by context and model capabilities.

May 8, 2026PLoS ONEOpen Access

Propensity to trust in Large Language Models

View Full Paper

Ask AI

Bookmark

This research aims to investigate the propensity to trust in large language models (LLMs) and understand how this trust is influenced by context and model capabilities.
Evaluated nineteen LLMs using a psychological self-report scale and a linguistic simulation framework.
Collected data from trust-related decisions to analyze the interaction between baseline delegation tendencies and model capabilities.
Conducted ablation studies to assess the impact of task-specific memory mechanisms on integrating trust cues.
Questionnaire showed uniformly high propensity to trust across models, indicating sycophantic response patterns.
Simulation framework revealed significant variations in models' trust behavior; high-capacity models adjust decisions based on trustworthiness cues.
Models like Llama-2-7B displayed stable delegation patterns and systematic over-entrustment, while more capable models modulated their baseline tendencies.

Discussion

No takes yet. Share an insight, caveat, or question.

Implication

Randomized trial examines trust behavior in large language models, suggesting implications for model trustworthiness assessment.

Cite This Study

Alice Plebe (2026) studied this question.

synapsesocial.com/papers/69fd7e23bfa21ec5bbf065d0 https://doi.org/10.1371/journal.pone.0347328

View Full Paper

Ask AI

Bookmark

Propensity to trust in Large Language Models

Key Points

Discussion

Implication

Cite This Study

Propensity to trust in Large Language Models

Key Points

Discussion

Implication

Cite This Study

Also Consider

Also Consider