Do I Believe My Robot is Really Sorry? Mind Perception’s Impact on Trust Repair in Robots

This post summarizes the paper “The Theory of Mind and Human-Robot Trust Repair,” which was co-authored by Connor Esterwood, and Lionel P. Robert Jr. This paper has been published in the Scientific Reports. You may access the paper here.

In our ever-evolving world, where robots are becoming integral parts of our lives, trust is a cornerstone of human-robot interactions. Just like humans, robots can err, and these mistakes can lead to a decline in trust. However, there’s a fascinating aspect to this dynamic — the potential for robots to repair our trust in them. This post delves into a study exploring the effectiveness of trust repair strategies in the context of human-robot interaction and how our perception of a robot’s mind plays a pivotal role.

Navigating Trust in a Robotic World

Picture this: you’re working alongside a robot in a warehouse, picking and loading boxes. Over time, the robot makes three mistakes, leaving you questioning your trust in its abilities. But here’s where it gets intriguing — the robot responds to each blunder with one of three strategies: a promise to improve, a denial of wrongdoing, or a heartfelt apology.

This scenario was at the heart of a study conducted with 400 participants recruited. The goal? To unravel the complex relationship between trust, repair strategies, and our perception of a robot’s mind.

The Three Repair Strategies

Before we dive into the study’s findings, let’s explore the three trust repair strategies:

Apology: The robot acknowledges its mistake and extends a sincere apology. This strategy relies on the belief that robots can experience emotions like humans.

Denial: The robot firmly denies any wrongdoing, suggesting that the mistake was a one-time anomaly. This approach might be more effective when humans perceive robots as intentional beings.

Promise: The robot commits to improving and ensuring that such mistakes won’t recur. This strategy hinges on rebuilding trust through demonstrated competence.

Mind Perception and Trust Repair

Now, here’s the crux of the matter: how our perception of a robot’s mind influences the effectiveness of these repair strategies. When humans attribute emotions to robots, apologies tend to be more impactful. Conversely, if we see robots as intentional actors, denials might resonate better.

The study unfolded in the context of the warehouse task, with participants rating their trust in the robot both before and after each mistake. The results were enlightening.

The Power of Mind Perception

Overall, the study underscored the significance of individual differences in mind perception. Apologies and denials between humans and robots are not one-size-fits-all. Instead, they depend on our beliefs about a robot’s mental capacities.

When participants perceived the robot as capable of experiencing emotions, heartfelt apologies led to substantial increases in trust. In contrast, for those who believed the robot possessed intentionality, firm denials were more effective in rebuilding trust.

Implications for Human-Robot Interaction

These findings carry substantial implications for the burgeoning field of human-robot interaction. They remind us that trust is not solely influenced by a robot’s actions but also by our perception of its inner workings. Implementing effective trust repair strategies hinges on understanding these intricate nuances.

As robots become increasingly integrated into our lives, these insights become more than mere academic curiosity. They offer a roadmap for developers, designers, and researchers to navigate the delicate terrain of human-robot trust. By tailoring repair strategies to align with our perception of a robot’s mind, we can nurture trust in robots and foster more harmonious human-robot relationships.

To learn more about this work and engage with MAVRIC, take a look at the paper here, and find us on Twitter (X), on the web, and in person at the University of Michigan Robotics. We welcome collaborators and interest from diverse disciplines. Hope to talk to you soon.