When the EU's AI Act partially took effect in August 2024, it was hailed as the first global framework to govern artificial intelligence. Within weeks, legislators introduced a rapid amendment package aimed at tightening definitions, expanding the scope to cover emerging models, and increasing penalties for non‑compliance.

The revision shifts several high‑risk categories, adds mandatory documentation for foundation models, and grants the European Commission broader enforcement powers. Critics argue the changes could raise compliance costs for tech firms operating across the bloc, while industry groups warn that uncertainty may delay investment in AI research.

By extending the Act's reach and sharpening enforcement, the EU signals its intent to set a de‑facto standard for AI governance worldwide. Companies will need to audit their AI pipelines now to avoid fines and market restrictions.