Human perceptions of social robot deception behaviors: an exploratory analysis

Area of Science:

• Human-Robot Interaction
• AI Ethics
• Social Robotics

Background:

• Robots are increasingly integrated into social environments, necessitating adherence to human social norms.
• Robot norm violations and deception present ethical challenges in human-robot interactions.

Purpose of the Study:

• To investigate how humans perceive and evaluate different types of robot deception.
• To analyze human responses to external state, hidden state, and superficial state deception behaviors in robots.

Main Methods:

• 498 participants read one of three vignettes describing different robot deception types.
• Quantitative and qualitative measures assessed perceptions of deceptiveness, approval, justification, and involvement of other agents.

Main Results:

• Hidden state deception was rated as most deceptive and least approved.
• External state and superficial state deception were perceived similarly, but only external state deception was generally approved.
• Participants often attributed hidden state deception to external agents, not the robot.

Conclusions:

• Humans distinguish between different types of robot deception, evaluating them differently in terms of deceptiveness and approval.
• Perception of deception, particularly hidden state deception, can involve attributing agency to robot designers rather than the robot itself.