PauseAI Hosts EU Conference on AI Security and the Control Problem
Why It Matters
The discussion marks a shift from focusing on algorithmic bias to addressing existential 'control' risks at a high legislative level within the EU. It signals increasing political pressure to regulate the speed of AI development beyond current AI Act frameworks.
Key Points
- MEP Ondřej Kolář called for regulation specifically targeting the pace of AI development to prevent harm.
- The conference focused on 'the control problem,' a technical safety concern regarding the alignment of advanced AI systems.
- PauseAI successfully moved existential risk discussions from fringe forums to the European Parliament halls.
- The event highlighted the dual-use nature of AI, balancing its medical potential against global security threats.
European Parliament Member Ondřej Kolář opened a significant conference titled 'Beyond the AI Act: Global security & the control problem' on February 23, 2026. Organized by the advocacy group PauseAI, the event gathered policymakers and safety researchers to discuss the dual-use nature of artificial intelligence. Kolář emphasized that while AI offers transformative benefits for medicine and research, the lack of pace-based regulation poses severe risks to global stability. He specifically addressed the 'control problem,' questioning whether current frameworks are sufficient to prevent catastrophic outcomes from autonomous systems. The conference reflects a growing movement within the European Union to reconsider safety thresholds as capability gains accelerate faster than previously anticipated by the AI Act.
At a conference hosted by PauseAI, EU lawmaker Ondřej Kolář compared AI to a powerful tool that is getting too fast for us to handle. Think of it like a sports car that keeps gaining speed while we are still trying to figure out if the brakes work. While he acknowledged that AI can help cure diseases, he warned that without slowing down and fixing the 'control problem'—the risk of losing grip on how these systems behave—we might face some truly terrible consequences. It is a sign that leaders are getting nervous about just how quickly this tech is moving.
Sides
Critics
Argues that current AI development is too fast and requires pace-based regulation to ensure global security.
Advocates for a global moratorium or slowdown on advanced AI development until safety can be guaranteed.
Defenders
No defenders identified
Neutral
Currently overseeing the implementation of the AI Act while debating further safety measures.
Noise Level
Forecast
Legislative proposals for 'compute caps' or development pauses are likely to be introduced in the EU Parliament as a follow-up to this conference. We will likely see more friction between safety advocates and tech lobbyists over the definition of 'controllable' AI.
Based on current signals. Events may develop differently.
Timeline
Conference Opens in EU Parliament
MEP Ondřej Kolář delivers the opening remarks at the 'Beyond the AI Act' conference organized by PauseAI.
Join the Discussion
Discuss this story
Community comments coming in a future update
Be the first to share your perspective. Subscribe to comment.