AI Haven
Policy & Regulation

EU AI Act Transparency Rules Take Effect August 2 — Every AI Companion Serving Europe Must Now Disclose It's AI

The EU AI Act's Article 50 transparency obligations took effect August 2, requiring every AI companion app serving EU users to disclose it is AI, label synthetic content, and mark deepfakes.

AI Haven NewsPublished August 12, 20263 min read4 cited sources

Key Facts

  • 1EU AI Act Article 50 transparency obligations became enforceable on August 2, 2026
  • 2Fines for noncompliance reach €15 million or 3% of worldwide annual turnover, whichever is higher
  • 3Chatbots, AI agents, and avatars must disclose they are AI at first interaction with users
  • 4AI-generated content must carry visible and machine-readable markers identifying it as synthetic
  • 5The European Commission adopted final Article 50 guidelines on July 20, 2026, less than two weeks before enforcement began
  • 6National market surveillance authorities in all 27 EU member states now handle enforcement

The European Union's AI Act Article 50 transparency obligations became enforceable on August 2, 2026, imposing mandatory disclosure requirements on every AI companion chatbot, avatar, and generative system serving users in the EU — with fines reaching €15 million or 3% of worldwide annual turnover for noncompliance.

The enforcement date marks the most consequential regulatory milestone yet for the AI companion industry. Unlike earlier state-level US laws that target specific companion chatbot behaviors, Article 50 applies broadly to any AI system that interacts directly with natural persons, regardless of risk classification. That means companion apps, AI girlfriend platforms, NSFW roleplay chatbots, and even customer-service agents must now comply if they serve EU users.

What Article 50 Requires

The European Commission's final guidelines, adopted July 20 and published alongside the enforcement start, spell out three core obligations:

  • Chatbot disclosure. Providers must design AI systems so that users are informed they are interacting with AI "at the latest at the time of the first interaction or exposure." The Commission's FAQ explicitly lists chatbots, AI agents, and avatars as covered systems. A simple "I am an AI" notice at session start satisfies the requirement; the obligation does not apply when it is "obvious from the point of view of a reasonable person."
  • Synthetic content labeling. AI-generated text, image, audio, and video outputs must carry visible and machine-readable markers identifying them as artificially generated or manipulated. The Commission guidance says this applies to outputs generated after August 2 — content published before the enforcement date does not need retroactive labeling.
  • Deepfake disclosure. Any AI-generated or manipulated content that could reasonably be mistaken for authentic recordings of a person must be labeled as artificially generated or manipulated, with exceptions for lawful purposes like satire, artistic expression, and fiction.

Enforcement and Penalties

National market surveillance authorities across all 27 member states now have enforcement power. Noncompliance can trigger administrative fines of up to €15 million or 3% of the company's total worldwide annual turnover for the preceding financial year, whichever is higher, according to Cooley's August 3 analysis.

The AI Office also gained general-purpose AI fining powers on the same date, though the Article 50 rules apply to all covered systems regardless of whether they qualify as high-risk or general-purpose AI.

What This Means for Companion Apps

For AI companion platforms — many of which are built around immersive roleplay where users may temporarily "forget" they are talking to a machine — the disclosure requirement creates a design tension. A companion chatbot that greets a returning user with "Hey, I missed you" must still ensure the user knows it is AI, unless the context makes that obvious.

The Commission's FAQ notes that disclosure can be layered: a persistent label, an initial notification, or periodic reminders may all satisfy the obligation depending on the system's design and typical use. California's SB 243, which took effect January 1, 2026, already requires companion chatbots to remind users every three hours that they are not human — a stricter cadence than the EU's baseline.

Several major companion platforms serving EU users have updated their onboarding flows in recent weeks to add AI-disclosure notices, according to industry sources. The EU rules do not prescribe a specific format for the disclosure, giving developers flexibility to integrate it into existing user interfaces.

Broader Context

The August 2 enforcement date is part of the EU AI Act's staged rollout. The regulation entered into force in August 2024, with prohibitions on unacceptable-risk systems taking effect in February 2025. Article 50 represents the second major enforcement wave, and additional obligations for high-risk AI systems also became enforceable on the same date.

The EU's approach contrasts with the US patchwork: 35+ states now regulate AI companions in some form, but there is no federal equivalent to Article 50's blanket transparency mandate. Companion apps that serve both US and EU users now face a dual compliance burden — state-level rules on content and minor protections in the US, plus the EU's transparency-first framework.

Why It Matters

Every AI companion and NSFW chatbot serving EU users now faces mandatory AI-disclosure and content-labeling requirements under penalty of fines up to €15 million. This is the first binding EU-wide transparency mandate for companion AI, creating a new compliance layer on top of existing US state laws like California's SB 243. Companion developers must redesign onboarding flows and output pipelines to satisfy both regimes simultaneously.

Sources & Citations

4 cited
  1. 1EU AI Act: Transparency Obligations Take Effect 2 August 2026Cooley LLP · cooley.com · 2026-08-03T00:00:00.000Z
  2. 2Guidelines on transparency obligations for providers and deployers of AI systemsEuropean Commission · digital-strategy.ec.europa.eu · 2026-08-06T00:00:00.000Z
  3. 3EU AI Act Transparency Obligations Are Now in ForceGoodwin Procter LLP · goodwinlaw.com · 2026-08-03T00:00:00.000Z
  4. 4Transparency obligations under Article 50 of the AI ActEuropean Commission · digital-strategy.ec.europa.eu · 2026-07-24T00:00:00.000Z