We encourage participants to share their research ideas, questions, and opinions with regard to the following themes:

  • Discovering and Identifying Cognitive Biases: We would like to explore mechanisms and components of AI systems that trigger cognitive biases in their users. In what human-AI collaboration scenarios are cognitive biases involved? Recent research has explored a diverse set of cognitive biases people follow when interacting with explainable AI systems.
  • Modelling and Quantifying Cognitive Biases: An important step towards bias mitigation is to model cognitive biases and measure their extent. However, since users are often unaware of their cognitive biases, it is challenging to know whether cognitive biases are manifesting themselves. Recent research has proposed mathematical frameworks to model cognitive biases. Moreover, some works have explored methods to reliably quantify cognitive biases in-situ using a variety of physiological sensors.
  • Novel Approaches to Mitigate Cognitive Biases: We would like to explore novel methods to mitigate the negative effects of cognitive biases in human-AI collaboration. Existing approaches include nudging, i.e., changing the choice environment, boosting, i.e., fostering metacognitive skills in people, and designing decision support systems that help users make effective and accurate decisions. We seek to discuss the shortcomings and limitations of existing debiasing approaches and develop future directions.
  • Application Scenarios of Cognitive Biases: While it is known that cognitive biases negatively affect human decision-making, we would like to explore the use of cognitive biases for the greater good. Can we imagine scenarios in which cognitive biases actually benefit human-AI collaboration?
  • Impact of the Bias Mitigation: We seek to explore how bias identification and mitigation strategies can positively and negatively impact AI systems and their users. What benefits do people get if their biases are mitigated? Do we really need to eliminate biases? Is there an alternative way to support human decision-making? Recent research has shown that some debiasing interventions like nudges can harm user autonomy or slow down the interaction.
  • Case Studies of Cognitive Biases in Human-AI Collaboration: Presentation of concrete cases where the prevalence, mitigation, and utilisation of cognitive biases in human-AI collaboration have been investigated.

Goals and Outcomes

We aim to foster a discussion about ongoing work on cognitive biases in HCI, provide a common platform to revisit the current research, and establish a research agenda for understanding, quantifying, mitigating, and utilising cognitive biases. Ultimately, we seek to form a research community that works towards the design of Bias-Aware Systems, defined as computing systems that take into account the cognitive biases of their users. Through creating this community, we aim to build a shared understanding of cognitive biases and methods to measure, utilise, and mitigate their effects. We expect that discussions in this workshop lead to fruitful collaborations that leverage our understanding of cognitive biases in users.