Low-Moral Actions by Malicious Anonymous Operators of Avatar Robots

This research was conducted during my master’s studies at the Human-Robot Interaction Lab at Kyoto University. You can read the manuscript titled Investigation of Low-Moral Actions by Malicious Anonymous Operators of Avatar Robots here.

Imagine a world where people use robots to visit different places and countries. Unlike screens on wheels (telepresence robots), avatar robots can look like anything. One flaw of this is that it is not readily apparent to bystanders who the operator of the robot is. Malicious operators can use this anonymity to do low-moral actions, which we define as not only illegal acts but also acts that violate norms of society.

In (Shaheen et al., 2024), we looked at what low-moral actions were possible and what prevention mechanisms were applicable for each. We identified four categories and 15 subcategories of malicious acts that can be done when a robot is limited to locomotor movement and video feed from the robot’s eyes to the operator only. Before the workshops, participants experienced avatar robots by controlling both a simulated avatar and a real avatar as a malicious anonymous operator in a variety of situations. They also experienced sharing space with an avatar controlled by a malicious anonymous operator.

Participants in an activity replicating an art or poster exhibition (one of the two activities). The video shows one participant being forced to change their position by the avatar operated by another participant from a separate room.

Robot inhibits movement of the crowd in a big clothing store

Robot inhibits movement of an individual in a large hall in a mall.

Examples of participants doing low-moral actions in three of the five simulation environments. Left: Robot inhibits movement of the crowd in a big clothing store. Right: Robot inhibits movement of an individual in a large hall in a mall. The top-down view was hidden from the participants. They only saw from the Robot's point-of-view.

We conducted three hazard identification workshops with a total of 12 participants (6 male, 6 female) to brainstorm possible low-moral actions. Participants first experienced avatar robots by operating a real robot, a simulated avatar, and sharing space with an avatar robot. Activities included tasks like navigating through simulated environments such as shopping malls and convenience stores, where participants acted as malicious operators. This experience phase was critical for ideation, ensuring that participants understood the technology and its potential for misuse.

Through these workshops, we identified four main categories of low-moral actions:

Violate Privacy and Security: Actions such as eavesdropping or entering forbidden spaces.
Inhibit: Blocking access to resources or obstructing movement.
Annoy: Actions like unnatural movement or invading personal space.
Destroy or Hurt: Causing physical harm to individuals or damage to property.

Each category was further divided into subcategories, resulting in a comprehensive list of potential actions and their manifestations. For instance, “Inhibit Movement” included examples like blocking hallways, while “Annoy” covered actions such as chasing individuals or misleading them.

Avatar robots allow a teleoperator to interact with the people and environment of a remote place. Malicious operators can use this technology to perpetrate malicious or low-moral actions. In this study, we used hazard identification workshops to identify low-moral actions that are possible through the locomotor movement, cameras, and microphones of an avatar robot. We conducted three workshops, each with four potential future users of avatars, to brainstorm possible low-moral actions. As avatars are not yet widespread, we gave participants experience with this technology by having them control both a simulated avatar and a real avatar as a malicious anonymous operator in a variety of situations. They also experienced sharing space with an avatar controlled by a malicious anonymous operator. We categorized the ideas generated from the workshops using affinity diagram analysis and identified four major categories: violate privacy and security, inhibit, annoy, and destroy or hurt. We also identified subcategories for each. In the second half of this study, we discuss all low-moral action subcategories in terms of their detection, mitigation, and prevention by studying literature from autonomous, social, teleoperated, and telepresence robots as well as other fields where relevant.

References

2024