Docs

The EU AI Act — What You Need to Know

The EU AI Act is the first comprehensive AI regulation. It applies to AI systems placed on the EU market or affecting people in the EU. Non-EU companies serving EU users are in scope. This guide covers key provisions, risk categories, Article 50 transparency (enforcement August 2, 2026), penalties, and practical steps. Focus on implications, not legal text.

Key Provisions

Risk-based approach — Rules vary by risk level. Higher risk means stricter requirements.

Transparency — Many AI systems must disclose that they are AI. Article 50 details apply from August 2, 2026.

Prohibitions — Some uses are banned (e.g., social scoring, manipulative subliminal techniques).

Conformity — High-risk systems need conformity assessments, documentation, and human oversight.

Risk Categories

Unacceptable — Banned. Social scoring, manipulative AI, real-time biometric identification in public (with narrow exceptions).

High — Strict requirements. Conformity assessment, risk management, human oversight. Examples: critical infrastructure, education, employment, essential services.

Limited — Transparency obligations. Chatbots, deepfakes, emotion recognition. Must disclose AI use.

Minimal — Light touch. Most general-purpose AI and low-risk applications.

Who It Affects

Providers — Those who develop or place AI systems on the market. Obligations on design, documentation, and transparency.

Deployers — Those who use AI under their authority. Obligations on use, monitoring, and transparency.

Non-EU companies — If you serve EU users, you are in scope. Appoint an EU representative if required.

Article 50 Transparency (Enforcement August 2, 2026)

For providers — AI that interacts with people must clearly indicate AI involvement. Generative AI outputs should be marked in machine-readable format and detectable as AI-generated where feasible.

For deployers — Disclose deepfakes (AI-generated or manipulated image, audio, video). Disclose AI-generated text on matters of public interest. Inform people exposed to emotion recognition or biometric categorization.

Format — Information must be clear, distinct, and accessible at first interaction or exposure.

Exceptions — Law enforcement (authorized uses), artistic/creative/satirical works (limited disclosure).

Penalties

Fines can reach up to 35M EUR or 7% of global turnover for prohibited practices. Lower tiers for other violations. Proportional to size and severity.

What to Do Now

  1. Map your AI use — What systems do you deploy? What risk category?
  2. Transparency — For chatbots and customer-facing AI, ensure disclosure. Plan for Article 50 by August 2026.
  3. Documentation — High-risk systems need conformity and risk documentation. Start early.
  4. Vendor review — Check if your AI vendors are compliant. Ask for documentation.
  5. Legal counsel — Engage for your specific situation. This guide is not legal advice.

The Bottom Line

The EU AI Act is in force. Transparency obligations (Article 50) apply from August 2, 2026. Map your AI use, ensure disclosure for customer-facing AI, and prepare for higher obligations if you have high-risk systems. Non-EU companies serving EU users are in scope.

Related Reading