文章轉錄自Science Daily（點此連結）；文章主要論述：研究者使用機器人來研究什麼因素會使人信任或不信任陌生人，使用機器人在於研究者可以明確操控手勢、對話等各種變項，文章並未指出明確的信任建立機制，但最後敘述到，建立信任的過程是基於「雙方」的互動。信任一件物體（非人）比較簡單，還是信任一個人呢？譬如：你會信任某個品牌車子的性能，但卻可能不太信任駕駛者或製造商，所以，當人面對機器人時所建立的信任關係，真的是能類推至人和人之間的信任關係嗎？（這是大哉問）（有許多社會現象或電影故事描述，有人無法在現實中信任他人，因而寧願轉而信任起虛擬人物）（好吧，把「人」和「非人」作為二分切割，或許有人會不滿意，尤其在越來越多人工器官置於人類身體的科技時代下）；如文中所提，人類互動過程中會無意識下「模仿」（mimic）對方的身體語言，而彼此信任的人其身體語言亦會共舞（dance），此概念即是諮商實務中會教導的mirror技巧。

Researchers Use Robot to Determine How Human Strangers Develop Trust

David DeSteno and MIT's Jin Joo Lee worked with Nexi the robot on the research project. (Credit: Photo by Mary Knox Merrill.)

ScienceDaily (July 10, 2010) — What can a wide-eyed, talking robot teach us about trust?

A lot, according to Northeastern psychology professor David DeSteno, and his colleagues, who are conducting innovative research to determine how humans decide to trust strangers -- and if those decisions are accurate.

The interdisciplinary research project, funded by the National Science Foundation (NSF), is being conducted in collaboration with Cynthia Breazeal, director of the MIT Media Lab's Personal Robots Group, Robert Frank, an economist, and David Pizarro, a psychologist, both from Cornell.

The researchers are examining whether nonverbal cues and gestures could affect our trustworthiness judgments. "People tend to mimic each other's body language," said DeSteno, "which might help them develop intuitions about what other people are feeling -- intuitions about whether they'll treat them fairly."

This project tests their theories by having humans interact with the social robot, Nexi, in an attempt to judge her trustworthiness. Unbeknownst to participants, Nexi has been programmed to make gestures while speaking with selected participants -- gestures that the team hypothesizes could determine whether or not she's deemed trustworthy.

"Using a humanoid robot whose every expression and gesture we can control will allow us to better identify the exact cues and psychological processes that underlie humans' ability to accurately predict if a stranger is trustworthy," said DeSteno.

During the first part of the experiment, Nexi makes small talk with her human counterpart for 10 minutes, asking and answering questions about topics such as traveling, where they are from and what they like most about living in Boston.

"The goal was to simulate a normal conversation with accompanying movements to see what the mind would intuitively glean about the trustworthiness of another," said DeSteno.

The participants then play an economic game called "Give Some," which asks them to determine how much money Nexi might give them at the expense of her individual profit. Simultaneously, they decide how much, if any, they'll give to Nexi. The rules of the game allow for two distinct outcomes: higher individual profit for one and loss for the other, or relatively smaller and equal profits for both partners.

"Trust might not be determined by one isolated gesture, but rather a 'dance' that happens between the strangers, which leads them to trust or not trust the other," said DeSteno, who, with his colleagues, will continue testing their theories by seeing if Nexi can be taught to predict the trustworthiness of human partners.