EU AI Act Extraterritoriality for Advertising
Tags: Regulatory
TL;DR
- Recital 29 clarifies that legitimate commercial practices like advertising aren’t inherently harmful manipulative AI.
- Article 2(1)(c) extends scope outside the EU if an AI system produces outputs within the EU.
- High‑risk systems face strict duties (transparency, protection, human oversight); typical advertising use cases often won’t be high‑risk unless they affect consequential decisions.
Why it matters for HK marketers: If your campaigns or tools output to EU users, EU AI Act obligations can reach you regardless of where you operate.
What the playbook highlights
- Recital 29 signal: Advertising that complies with applicable law isn’t by itself a prohibited manipulative practice.
- Extraterritorial reach: Outputs in the EU can trigger coverage even for non‑EU organizations (Art. 2(1)(c)).
- High‑risk focus: Systems tied to biometric processing, critical infrastructure, law enforcement, or access to essential services carry added obligations.
Advertising risk posture
- Most ad use cases likely fall outside current high‑risk definitions unless they influence access to housing, employment, healthcare, or similar.
- Still, assess sector laws (e.g., health privacy) and audience profiling impacts.
So what for marketers
When targeting EU audiences, document AI use cases, data flows, and human oversight. Treat health and biometric inferences as high‑scrutiny areas even if not formally high‑risk.
← Back to Knowledge Base