• Speaker: Marc Rotenberg, Center for AI and Digital Policy, USA
• Moderator: Anna Bon, UV Amsterdam, The Netherlands

About the Event

November 25, 2025
5:00 – 6:00 PM
(17:00) CEST

We are looking forward to seeing you:

Participate via Zoom (Password: 0dzqxqiy).
The talk will also be live-streamed and recorded on the DIGHUM YouTube Channel.

Abstract

Around the world, policymakers and researchers are converging on a simple conclusion: the future of trustworthy AI depends on establishing clear red lines for systems that undermine human rights, safety, or democratic institutions. Leading AI scientists—including Russell, Hinton, and Bengio—have warned that some advanced systems exhibit behaviors developers cannot fully predict or control, underscoring the need for enforceable limits. Human-rights bodies have reached similar judgments: the UN High Commissioner for Human Rights and UNESCO have called for bans on AI uses such as mass surveillance and social scoring. Recent frameworks—from the EU AI Act to the Council of Europe AI Convention—translate these concerns into concrete prohibitions. Building on earlier principles such as the Termination Obligation in the Universal Guidelines for AI, the new AI Red Line Initiative seeks to amplify these emerging prohibitions. The evidence is now clear: well-defined red lines do help. They protect fundamental rights, provide operational clarity for developers, and create space for beneficial innovation within boundaries that preserve human control.

Slides

Slides will be available for download after the lecture.

Video

Recording will be available after the lecture.