If They Disagree, Will You Conform? Exploring the Role of Robots' Value Awareness in a Decision-Making Task

TL;DR

This study asks whether robots with value-awareness can sway human decisions in value-labeled image tasks. It uses two Furhat robots—one value-aware (VA) and one non-value-aware (NVA)—in an Asch-like, multi-party setup, measuring explicit choices, gaze behavior, and moral- Foundations-based perceptions. Results show that explicit conformity to robot opinions is modest (about 25% when both robots disagree), while participants gaze more at the VA robot and rate it higher on loyalty and purity, indicating nuanced social influence beyond explicit decisions. The work highlights safety and design implications for deploying value-aware social robots, suggesting they can foster reflection in ambiguous decisions while also presenting potential for manipulation if misused. Overall, the findings advance understanding of how value alignment affects trust, engagement, and group dynamics in human-robot teams.

Abstract

This study investigates whether the opinions of robotic agents can influence human decision-making when robots display value awareness (i.e., the capability of understanding human preferences and prioritizing them in decision-making). We designed an experiment in which participants interacted with two Furhat robots - one programmed to be Value-Aware and the other Non-Value-Aware - during a labeling task for images representing human values. Results indicate that participants distinguished the Value-Aware robot from the Non-Value-Aware one. Although their explicit choices did not indicate a clear preference for one robot over the other, participants directed their gaze more toward the Value-Aware robot. Additionally, the Value-Aware robot was perceived as more loyal, suggesting that value awareness in a social robot may enhance its perceived commitment to the group. Finally, when both robots disagreed with the participant, conformity occurred in about one out of four trials, and participants took longer to confirm their responses, suggesting that two robots expressing dissent may introduce hesitation in decision-making. On one hand, this highlights the potential risk that robots, if misused, could manipulate users for unethical purposes. On the other hand, it reinforces the idea that social robots could encourage reflection in ambiguous situations and help users avoid scams.

Figures (5)

• Figure 1: An item of the stimuli validation survey. Participants use the cursor to indicate on a Likert scale which of the two words is more closely associated with the image set.
• Figure 2: A photo of the experimental setup taken during a session, with the participant’s consent. In particular, the photo was taken during the Description task: the participant is describing the picture on the screen.
• Figure 3: Flow chart of the experimental procedure.
• Figure 4: Percentages of participants' changed answers for different combinations of VA and NVA Furhat in disagreement with them, during the Images-Keyword Association.
• Figure 6: Histogram showing percentages of participants' gaze toward the screen and the robots during the Description Task. The bars are split into three sets, according to the agent describing the picture. Error bars represent the standard error of the mean.