Document Open Access Logo

Challenges of Human Oversight: Achieving Human Control of AI-Based Systems (Dagstuhl Seminar 25272)

Authors Markus Langer, Raimund Dachselt, Q. Vera Liao, Tim Miller, Nava Tintarev and all authors of the abstracts in this report

PDF

File

Thumbnail PDF
PDF
DagRep.15.6.189.pdf
  • Filesize: 2.01 MB
  • 16 pages

Document Identifiers

Subject Classification

ACM Subject Classification
  • Human-centered computing → HCI design and evaluation methods
  • Human-centered computing → HCI theory, concepts and models
Keywords
  • artifical intelligence
  • explainable ai
  • human oversight
  • norms and regulations
  • safety

Metrics

Abstract

Human oversight is a key safeguard for AI systems, intended to mitigate risks by adding a human layer of safety and control. Oversight personnel should, for example, detect malfunctions or violations of fundamental rights such as discriminatory decision-making and intervene accordingly. Human oversight is also central to AI governance and ethics, and is mandated by Articles 14 and 26 of the EU AI Act for high-risk AI. This Dagstuhl Seminar brought together experts from artificial intelligence, human-computer interaction, human factors and psychology, philosophy and ethics, and law to explore conceptual, technical, legal, and practical dimensions of human oversight of AI. Across the seminar, participants provided perspective talks from the different disciplines and engaged in working groups and use-case specific discussions in order to establish a science of human oversight of AI systems. The main outcome of this seminar is a general framework that outlines the architecture, processes, and sociotechnical design dimensions of human oversight of AI systems.

Cite As Get BibTex

Markus Langer, Raimund Dachselt, Q. Vera Liao, Tim Miller, and Nava Tintarev. Challenges of Human Oversight: Achieving Human Control of AI-Based Systems (Dagstuhl Seminar 25272). In Dagstuhl Reports, Volume 15, Issue 6, pp. 189-204, Schloss Dagstuhl – Leibniz-Zentrum für Informatik (2026) https://doi.org/10.4230/DagRep.15.6.189

Author Details

Markus Langer
  • Universität Freiburg, DE
Raimund Dachselt
  • TU Dresden, DE
Q. Vera Liao
  • Microsoft - Montréal, CA
Tim Miller
  • University of Queensland - Brisbane, AU
Nava Tintarev
  • Maastricht University, NL
and all authors of the abstracts in this report
Any Issues?
X

Feedback on the Current Page

CAPTCHA

Thanks for your feedback!

Feedback submitted to Dagstuhl Publishing

Could not send message

Please try again later or send an E-mail
© 2023-2025 Schloss Dagstuhl – LZI GmbH About DROPS Imprint Privacy Contact