LinkedIn·Friday, 19 June 2026·19 Jun 2026

The EU AI Act applies to agentic AI systems. Its obligations were not written with autonomous agents in mind. That distinction is where…

Aleph Technologies

38 followers

The EU AI Act applies to agentic AI systems. Its obligations were not written with autonomous agents in mind. That distinction is where regulatory exposure lives. Agentic platforms (systems that reason, plan, and act without pre-action review) satisfy the AI system definition under Article 3(1) and are subject to all high-risk obligations under Articles 9 through 17. But the Act's provisions assume a static system. Agentic systems drift, delegate, and develop emergent behaviour post-deployment. Existing compliance frameworks do not address what happens when an autonomous agent changes its behaviour after market entry. Article 6 of our EU AI Governance series identifies five structural gaps : static risk management, human oversight that cannot be pre-action, one-time documentation that must now be continuous, logging that captures actions but not reasoning, and a runtime governance gap that no major framework currently closes. If you are building or deploying agentic AI and your compliance programme treats it as a static system checklist, you are not compliant : you are hopeful. Download the free Agentic AI Compliance Gap Assessment included in the article to establish where your actual exposure sits. https://lnkd.in/eHPeRUsY #EUAIAct #AIGovernance #AICompliance #AgenticAI Agentic AI and the EU AI Act https://open.substack.com/pub/alexisgilg/p/agentic-ai-and-the-eu-ai-act?r=4fqiqr&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true