본문으로 건너뛰기
All news

EU AI Act Enforcement Begins: Chatbots Must Disclose, Deepfakes Must Be Labeled

Summary: Europe's AI Act moved from policy to enforcement on August 2, flipping on transparency rules, high-risk AI obligations, and the EU AI Office's investigative powers in a single day.

Key Facts

  • Disclosure mandate: Interactive AI systems — chatbots, voice assistants — must now inform EU users they are speaking with an AI, not a human.
  • Watermarking required: Deepfakes and AI-generated or altered content must carry machine-readable markers so they can be detected automatically.
  • High-risk obligations live: Requirements for AI deployed in hiring, credit scoring, healthcare, and law enforcement officially took effect on the same date.
  • Penalties: Violations risk fines up to €15 million or 3% of worldwide annual turnover, whichever is higher.

Why It Matters

This is the first time a major jurisdiction has switched on binding, enforceable rules for general-purpose AI models at scale. The EU AI Office can now demand documentation, access models, order corrective measures, and levy fines against providers like OpenAI, Anthropic, and Google — all of whom serve European users. The practical compliance burden falls hardest on companies whose products blur the line between AI and human interaction, making this a watershed moment for how AI is deployed globally.

Further Reading

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.