Employment outcomes for autistic ¹ individuals are often poorer relative to their neurotypical (NT) peers, resulting in a greater need for other forms of financial and social support. While a great deal of work has focused on developing interventions for autistic children, relatively less attention has been paid to directly addressing the employment challenges faced by autistic adults. One key impediment to autistic individuals securing employment is the job interview. Autistic individuals often experience anxiety in interview situations, particularly with open-ended questions and unexpected interruptions. They also exhibit atypical gaze patterns that may be perceived as, but not necessarily indicative of, disinterest or inattention. In response, we developed a closed-loop adaptive virtual reality (VR)–based job interview training platform, which we have named Career Interview Readiness in VR (CIRVR). CIRVR is designed to provide an engaging, adaptive, and individualized experience to practice and refine interviewing skills in a less anxiety-inducing virtual context. CIRVR contains a real-time physiology-based stress detection module, as well as a real-time gaze detection module, to permit individualized adaptation. We also present the first prototype of the CIRVR Dashboard, which provides visualizations of data to help autistic individuals as well as potential employers and job coaches make sense of the data gathered from interview sessions. We conducted a feasibility study with 9 autistic and 8 NT individuals to assess the preliminary usability and feasibility of CIRVR. Results showed differences in perceived usability of the system between autistic and NT participants, and higher levels of stress in autistic individuals during interviews. Participants across both groups reported satisfaction with CIRVR and the structure of the interview. These findings and feedback will support future work in improving CIRVR’s features in hopes for it to be a valuable tool to support autistic job candidates as well as their potential employers.

Autistic individuals face difficulties in finding and maintaining employment, and studies have shown that the job interview is often a significant barrier to obtaining employment. Prior computer-based job interview training interventions for autistic individuals have been associated with better interview outcomes. These previous interventions, however, do not leverage the use of multimodal data that could give insight into the emotional underpinnings of autistic individuals’ challenges in job interviews. In this article, the authors present the design of a novel multimodal job interview training platform called CIRVR that simulates job interviews through spoken interaction and collects eye gaze, facial expressions, and physiological responses of the participants to understand their stress response and their affective state. Results from a feasibility study with 23 autistic participants who interacted with CIRVR are presented. In addition, qualitative feedback was gathered from stakeholders on visualizations of data on CIRVR’s visualization tool called the Dashboard. The data gathered indicate the potential of CIRVR along with the Dashboard to be used in the creation of individualized job interview training of autistic individuals.

Autistic adults possess many skills sought by employers, but may be at a disadvantage in the workplace if social-communication differences negatively impact teamwork. We present a novel collaborative virtual reality (VR)-based activities simulator, called ViRCAS, that allows autistic and neurotypical adults to work together in a shared virtual space, offering the chance to practice teamwork and assess progress. ViRCAS has three main contributions: 1) a new collaborative teamwork skills practice platform; 2) a stakeholder-driven collaborative task set with embedded collaboration strategies; and 3) a framework for multimodal data analysis to assess skills. Our feasibility study with 12 participant pairs showed preliminary acceptance of ViRCAS, a positive impact of the collaborative tasks on supported teamwork skills practice for autistic and neurotypical individuals, and promising potential to quantitatively assess collaboration through multimodal data analysis. The current work paves the way for longitudinal studies that will assess whether the collaborative teamwork skill practice that ViRCAS provides also contributes towards improved task performance.

The interview process has been identified as one of the major barriers to employment of autistic individuals, which contributes to the staggering rate of under and unemployment of autistic adults. Decreasing stress during the interview has been shown to improve interview performance. However, in order to effectively provide insights on stress to both interviewees and interviewers, it is necessary to first effectively measure stress. This work explores physiological stress detection through wearable sensing as a means of obtaining quantitative stress measures from young autistic adults undergoing a virtual simulated interview using supervised machine learning techniques. Several supervised learning models were explored and it was found that Elastic Net Regression had the best accuracy with individual models with an accuracy of 84.8% while Support Vector Regression models evaluated with leave-one-out cross validation had a group accuracy of 75.4%. The predictions from the stress model were used with data visualization techniques in order to provide insights on the interview process from both a group and individual viewpoint, showing that stress trends can be found and evaluated using the stress model. This work also addresses a major gap in physiological stress detection literature by presenting a novel dataset of physiological data and ground truth labels for 15 autistic young adults undergoing a simulated interview.

Computer-based job interview training, including virtual reality (VR) simulations, have gained popularity in recent years to support and aid autistic individuals, who face significant challenges and barriers in finding and maintaining employment. Although popular, these training systems often fail to resemble the complexity and dynamism of the employment interview, as the dialogue management for the virtual conversation agent either relies on choosing from a menu of prespecified answers, or dialogue processing is based on keyword extraction from the transcribed speech of the interviewee, which depends on the interview script. We address this limitation through automated dialogue act classification via transfer learning. This allows for recognizing intent from user speech, independent of the domain of the interview. We also redress the lack of training data for a domain general job interview dialogue act classifier by providing an original dataset with responses to interview questions within a virtual job interview platform from 22 autistic participants. Participants’ responses to a customized interview script were transcribed to text and annotated according to a custom 13-class dialogue act scheme. The best classifier was a fine-tuned bidirectional encoder representations from transformers (BERT) model, with an f1-score of 87%.