Negotiation skills training intervention based on automated recognition of human emotion and nonverbal behaviour

Abstract

Overview: The current Ph.D. thesis investigates whether it is possible to train individuals in the skill of using social signals in interactive scenarios by using the outputs of automated emotion detection and recognition technologies, and whether this can result in improved performance. The specific focus of the research is on improving understanding of the extent to which people can act on social signals feedback to adapt their subsequent displays of social signals in a negotiation context. The findings have practical implications for recommendations for negotiation skills training practises in educational and operational settings. Aim: The goal of the Ph.D. thesis is to determine how new commercial off-the-shelf technological advances for the automated recognition of human emotion expression and nonverbal cues might be used to aid in the development of negotiation abilities. Method: The aim was addressed through a programme of empirical work conducted in three stages: (1) an exploratory research stage to establish how negotiation skills training is currently carried out within the UK MOD, (2.1) an intervention design stage to incorporate social signals feedback into negotiation skills training, (2.2) an experimental investigation stage (n=50) to determine the impact of social signals training intervention on negotiation outcomes, (3) stakeholder and military community evaluation stage to obtain feedback on the applicability of the semi-autonomous social signals training intervention to wider MOD. Results: The findings show that combining negotiation skills training with social signals feedback improves objectively evaluated joint gain outcomes for both negotiating parties. Additionally, the findings show that facial emotion expressions ('smile’, ‘joy, 'anger’, 'sadness’, 'disgust’, 'contempt’, and ‘fear'), ‘honest signals’ (mimicry, specifically 'posture mirroring'), and voice emotion expression of 'upset' are important for joint gain outcomes in bargaining scenarios. Finally, present findings suggest that social signals such as ‘honest signals’ (mimicry, activity, and influence), facial emotion expressions, and ‘extreme voice emotion’ can be trained and enhanced with individualised post-action social signals feedback. Conclusion: This Ph.D. thesis has effectively demonstrated the ability of social signals feedback to enhance negotiation outcomes when compared to traditional training methods. Given the exploratory character of the study, it is suggested that future research builds on the merits of the current study while addressing its shortcomings with a more focused approach to investigating social signals use in training contexts.