# EU AI Act Gets Teeth — Binding Transparency While Washington Stays Voluntary

Times of AI Desk · 2026-08-02 · Policy

[https://timesof.ai/2026/08/eu-ai-act-enforcement-begins-august-2](https://timesof.ai/2026/08/eu-ai-act-enforcement-begins-august-2)

> The European Commission’s AI Office and national authorities begin enforcing the AI Act’s remaining core rules — including transparency duties for chatbots and deepfakes and supervisory powers over GPAI models. Brussels becomes the world’s first comprehensive AI regulator with live penalty teeth; the U.S. same week is still building a voluntary, partly classified path.

**Enforceable** AI law changes product roadmaps, disclosure UX, watermarking pipelines, and legal budget — not just policy blogs. Europe chooses **binding, public statutory duties** the same week Washington finalizes a **voluntary, partly classified** pre-release review path.

**August 2, 2026** is the day the **EU Artificial Intelligence Act**’s main remaining obligations become **enforceable**. The **European Commission’s AI Office**, together with **Member State authorities**, can supervise and penalize non-compliance. The same day, **Article 50-style transparency rules** apply: interactive AI systems must disclose they are AI; **deepfakes** must be labelled; AI-generated or altered content needs **machine-readable** marks. The Commission had previewed the switch-on in a **31 July 2026** press package.

## What becomes live

| Surface | What changes (EC / digital-strategy primary) |
|---------|-----------------------------------------------|
| **Enforcement start** | AI Office + national authorities: implement, supervise, enforce |
| **GPAI models** | AI Office holds enforcement over **general-purpose AI** providers (docs, evaluation, corrective measures, fines) |
| **Transparency** | Chatbots/interactive systems must tell users they are AI; deepfakes labelled; AI content machine-readable marked |
| **Sandboxes** | Member States must have at least one **national AI regulatory sandbox** |
| **High-risk systems** | Bulk high-risk operator obligations apply (with staged exceptions still on the statutory calendar) |

The AI Act **entered into force 1 August 2024** and applied in **phases**. **2 August 2026** is the milestone that turns the Commission from rule-writer into day-to-day **market supervisor** for the most consequential AI obligations.

## Operational detail for labs and deployers

- **GPAI Code of Practice** and **GPAI provider guidelines** operationalize transparency, copyright, and systemic-risk duties for frontier-class models.
- A first list of **180+** organisations signed the **Code of Practice on transparency of AI-generated content** (Commission note in the 31 July package).
- Complaint and whistleblower channels for AI Act breaches and **downstream GPAI** issues are live on Commission digital-strategy surfaces.
- Fines for non-compliance can reach **up to 7% of global annual turnover** for the most serious breaches under the Act’s penalty architecture (as widely briefed in contemporaneous coverage).

U.S. and other non-EU labs shipping into the EU single market remain in scope when they place models or systems on the EU market — extraterritorial reach is the design, not a side effect.

## Transatlantic contrast (same week)

The EU switch-on lands as Washington finalizes a **voluntary** cyber-capability testing framework for frontier U.S. models (see August 3 White House coverage). Multinationals now run two compliance clocks in parallel.

## Limits

- First inspection waves and lead systemic-risk cases are ahead — this is the switch-on, not the first fine.
- Staged exceptions for some high-risk obligations remain on the statutory calendar.
- Whether open-weight providers treat EU marking rules as global defaults is an adoption question.

## Sources

- [European Commission digital strategy: Commission starts enforcing AI Act rules and new transparency requirements on 2 August (31 July 2026)](https://digital-strategy.ec.europa.eu/en/news/commission-starts-enforcing-ai-act-rules-and-new-transparency-requirements-2-august)
- [European Commission: Enforcement of the AI Act](https://digital-strategy.ec.europa.eu/en/policies/enforcement-ai-act)
- [European Commission: Regulatory framework for AI (AI Act overview)](https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai)
- [Euronews: EU rules on AI models become enforceable (August 2, 2026)](https://www.euronews.com/my-europe/2026/08/02/eu-rules-on-ai-models-become-enforceable-whats-going-to-change)