https://www.dagstuhl.de/20442

25. – 28. Oktober 2020, Dagstuhl-Seminar 20442

CANCELLED Detection and Design for Cognitive Biases in People and Computing Systems

Due to the Covid-19 pandemic, this seminar was cancelled.

Organisatoren

Andreas Dengel (DFKI – Kaiserslautern, DE)
Tilman Dingler (The University of Melbourne, AU)
Evangelos Karapanos (Cyprus University of Technology – Limassol, CY)
Koichi Kise (Osaka Prefecture University, JP)
Benjamin Tag (The University of Melbourne, AU)

Auskunft zu diesem Dagstuhl-Seminar erteilen

Jutka Gasiorowski zu administrativen Fragen

Shida Kunz zu wissenschaftlichen Fragen

Motivation

With social computing systems and algorithms having been shown to give rise to unintended consequences, one of the suspected catalysts is their ability to integrate, utilize, and reinforce people's inherent cognitive biases. Biases can be present in users, systems and their contents or applications: cognitive biases in users increase susceptibility to various types of manipulation, such as confirmation biases (predominantly seeking out information that confirms existing views) or cognitive dissonance (repudiation of information that does not fit into preconceived notions). These biases can be found in both individuals and organizations and are a crucial obstacle for rational, logical discussions of polarizing topics. Biases in systems are often consequences of their creators' inherent assumptions, the types of abstractions made during their creation or the specific data that fuels their algorithms. Especially systems that enable targeted distribution of information have recently come under increasing scrutiny. The effectiveness of micro-targeting and the successful distribution of falsified information through social media channels have led to increasing concerns about biased information consumption.

While biases have been researched in Psychology and Behavioral Economics for some time, their recent politicization has attracted attention and scrutiny in the public eye demanding a comprehensive research agenda that spans disciplines and application scenarios. This seminar, therefore, brings together designers, developers, and thinkers across disciplines to redefine computing systems by focusing on inherent biases in people, systems, and contents with the goal of working towards a research agenda to mitigate their effects. Three high-level questions will guide a mix of lightning talks, panel discussions, and workgroup sessions:

  1. How can cognitive biases in people, systems, and content be detected?
  2. How can such biases be made transparent, comprehensible, and actionable to users?
  3. What are practical techniques to mitigate the effects of biases in the shape of system interventions and user inoculation?

Over three days, an esteemed selection of participants will engage with these questions and re-think the notion of cognitive biases in computing systems and user interactions with particular regard to application scenarios in news media and decision-making support. With the proliferation of machine learning algorithms in our everyday life (e.g., for job applicant selection based on facial features or insurance claim assessments based on zip code and race), cases of biased decisions made by these algorithms have raised ethical concerns and demands for more transparent and explainable algorithms.

The goal of this Dagstuhl Seminar is to establish a comprehensive research agenda around cognitive biases, their detection, utilization, and possible mitigation as a response to emerging changes in societal and public discourse contributed to by recent technological and political developments. By looking at systems, users, and applications from an interdisciplinary perspective, we aim to produce blueprints for systems that provide transparency and contribute to advancing technology and media literacy, build critical thinking skills, and depolarize by design.

Motivation text license
  Creative Commons BY 3.0 DE
  Tilman Dingler and Benjamin Tag

Classification

  • Computers And Society
  • Human-Computer Interaction
  • Other Computer Science

Keywords

  • Cognition
  • Behaviour
  • Bias

Dokumentation

In der Reihe Dagstuhl Reports werden alle Dagstuhl-Seminare und Dagstuhl-Perspektiven-Workshops dokumentiert. Die Organisatoren stellen zusammen mit dem Collector des Seminars einen Bericht zusammen, der die Beiträge der Autoren zusammenfasst und um eine Zusammenfassung ergänzt.

 

Download Übersichtsflyer (PDF).

Publikationen

Es besteht weiterhin die Möglichkeit, eine umfassende Kollektion begutachteter Arbeiten in der Reihe Dagstuhl Follow-Ups zu publizieren.

Dagstuhl's Impact

Bitte informieren Sie uns, wenn eine Veröffentlichung ausgehend von
Ihrem Seminar entsteht. Derartige Veröffentlichungen werden von uns in der Rubrik Dagstuhl's Impact separat aufgelistet  und im Erdgeschoss der Bibliothek präsentiert.