본문으로 건너뛰기
All news

EU AI Act T-33 Days: Chatbot Disclosure and Deepfake Labels Become Law August 2

Summary: On August 2, 2026 — 33 days from today — the bulk of the EU AI Act takes effect, imposing disclosure, labeling, and audit requirements on anyone serving EU users.

Key Facts

  • Article 50 (Transparency): Chatbot operators must inform users they are interacting with AI; AI-generated images, audio, and video must be identifiably labeled; deepfakes published to inform the public on matters of public interest require clear, visible marking
  • High-Risk AI Systems: AI used in hiring, credit scoring, healthcare, law enforcement, migration, or critical infrastructure must complete conformity assessments, finalize technical documentation, affix CE marking, and register in the EU database — all by August 2
  • Penalties for non-compliance: up to €35M or 7% of global annual revenue, whichever is higher — more punitive than GDPR
  • A deadline extension for high-risk systems has been discussed but not enacted; enterprises should treat August 2 as the binding date

Why It Matters

Like GDPR, the EU AI Act applies wherever EU residents are served — meaning US and Korean companies are fully in scope. With 33 days left, any organization that hasn't started conformity work is running out of time. The transparency requirements alone affect virtually every deployed AI product.

Read More

뉴스레터 구독

무료 뉴스레터

매주 핵심 AI 소식, 한 번에 받기

쏟아지는 AI·LLM 뉴스 중 꼭 알아야 할 것만 골라 메일로 보내드려요. 뉴스레터 발송이 시작되면 구독자분들께 가장 먼저 보내드립니다.