Bots Against Bias

About

Bias has become a hot topic in robotics and beyond. Experts and the general public have come to recognize the key role that machines play in shaping human reactions and influencing human behavior. Whether purposefully or not, machines are being crafted in ways that take advantage of the natural cognitive pitfalls of the human brain and perpetuate skewed, if not harmful human biases. From the algorithms we are exposed to online to the way the robots we interact with are designed, we and the machines we make are at the heart of this crucial and pervasive challenge. We can no longer ignore it. Yet, how to effectively tackle this multifaceted challenge, one that is experienced on the user side as well as within us—designers, developers, engineers, and researchers—remains an open question.

One possibility is to flip the script by conscientiously (re)designing “bots”—robots, chatbots, intelligent agents, and other autonomous machines—that help us grapple with our biases and other cognitive pitfalls as well as augment metacognition and decision-making in our daily lives. To do this, we will need to bring together the latest advances in the fields of metacognition and robotics. We will need to integrate complementary, critical trajectories of research and methods in the human sciences as well as the science of machines. We will need to go beyond our disciplinary boundaries and develop new research questions, adapt our experimental designs, and shift our robotics praxis to establish this research field and make headway on this widespread problem.

Objectives

This new half-day workshop will bring together a diverse gathering of researchers and practitioners invested in confronting the challenge of bias and metacognition head-on through the rigorous and conscientious design of bots and their study. The intended audience and their backgrounds will be multidisciplinary: from cognitive robotics to social robotics, spanning the fields of human-robot interaction (HRI), human-agent interaction (HAI), and human-computer interaction (HCI), including expertise on machine learning (ML) and artificial intelligence (AI), and incorporating relevant topics from cognitive science, psychology, ergonomics/human factors (HF/E), education, neuroscience, and more. The goal of this inaugural workshop is twofold: (i) to establish a research agenda supplemented by start-up research, and (ii) to launch an international network of like-minded researchers and practitioners who will collaborate on next steps. The following themes will be the focus:

• Robotic aids for metacognition and decision-making
• Intelligent agents for supporting and enhancing human cognition
• Social and educational robots for anti-bias training and prevention
• Human behavior modelling and metacognition strategies
• Individual and social models of bias and metacognition
• Acceptance and trust of bots as metacognitive aids and partners
• Algorithmic bias and critical deconstructions of bias in robot design
• Explainable AI, ethical AI, and personalization of algorithms

How to Participate

This inaugural half-day workshop will be held online as part of RO-MAN 2021. We welcome original position papers, extended abstracts, or works-in-progress research articles on any of the following:

• Reviewing the state of the art on bots against bias, bots for metacognition, bots as aids in decision-making, and related topics
• Reviewing the pro-metacognition and anti-bias strategies and techniques that have or could be employed by bots and have demonstrated effectiveness for people
• Identifying or developing new critical frameworks and conceptual models for the design of these metacognition bots and their use by and with people
• Identifying situations, contexts, activities, and states of being in which bias plays a role or cognition can be compromised
• Developing experimental, exploratory, and qualitative research designs to evaluate the effectiveness and impact of these bots in and outside of the lab
• Developing requirements and design guidelines to jumpstart this trajectory of research
• Identifying critical barriers to multidisciplinary projects and collaborations
• Identifying personal and professional biases in bot design and evaluation
• Establishing new projects and collaborations to move the research area forward

Papers will be presented orally in 5-7 minutes with an accompanying 7-10 minutes of discussion. At least one author must attend the workshop. Registration at RO-MAN 2021 is NOT required.

We are aiming to provide attendees with the opportunity to extend their paper into a long article, to be included in a special issue of a topical journal (to be approved). As such, papers must not be under consideration elsewhere.