Beginner · Safety

EU AI Act Enforcement Phase

Visual diagram · (in preparation) · Math · (in preparation) · Worked example · 3 difficulty levels.

TL;DR. The period beginning August 2, 2026, when the European AI Act's obligations — including machine-readable marking of AI-generated content — became actively enforced.

Technical Definition

The period beginning August 2, 2026, when the European AI Act's obligations — including machine-readable marking of AI-generated content — became actively enforced.

How it works

The EU AI Act phased in over two years: prohibitions on unacceptable-risk systems arrived first, then obligations for general-purpose models, and on August 2, 2026 the bulk of the regulation entered active enforcement — including transparency duties, content-marking requirements, and rules for high-risk systems, backed by significant fines. For builders, the practical consequences are provenance watermarking on generated content, documented risk management for high-risk deployments, fundamental-rights impact assessments, and post-market monitoring. The enforcement wave directly drove 2026 investments in model-level watermarking, eval documentation, and governance tooling.

Related Concepts

• Model-Level Invisible Watermarking — A provenance technique — introduced by Anthropic for Claude — that embeds an unalterable, imperceptible watermark during token generation itself, surviving rewriting and copy-paste.
• AI Governance — The framework of policies, regulations, and principles designed to guide the development and deployment of AI technologies.