Judging One's Own or Another Person's Responsibility in Interactions With Automation

Area of Science:

• Human-computer interaction
• Cognitive psychology
• Automation and intelligent systems

Background:

• Human responsibility for outcomes is unclear in advanced automation.
• Actors and observers may have differing perceptions of responsibility.
• Understanding these differences is crucial for effective human-automation collaboration.

Purpose of the Study:

• To explore subjective assessments of human and automation capabilities.
• To investigate causal responsibility for outcomes in human-automation interaction.
• To compare perceptions between active users (actors) and passive observers.

Main Methods:

• A laboratory experiment involving pairs of participants (actor and observer).
• Actors performed a decision task with an automated system; observers watched.
• Perceptions of responsibility were compared across roles and two different system capabilities.

Main Results:

• Actors and observers similarly assessed system and human capabilities and comparative responsibility.
• Actors attributed adverse outcomes more to system characteristics than personal limitations.
• Observers underestimated system capabilities when assigning responsibility to actors.

Conclusions:

• When automation exceeds human ability, users may feel less responsible, potentially leading to over-interference.
• Outside observers may overestimate user contributions, assigning undue responsibility for negative outcomes.
• Calibrating subjective assessments through performance data can reduce biases in both users and observers.