Summary

The trilogue negotiations between the EU Parliament, member states, and the Commission on the Digital Omnibus ended early Wednesday morning without agreement. The main point of contention was the EPP group's demand – supported by German Chancellor Friedrich Merz – to exempt industrial AI applications from the scope of the AI Act. This blocks both the reform of high-risk AI regulations and the planned ban on sexualized deepfakes. The continuation of talks was postponed indefinitely.

People

Topics

  • EU regulation of artificial intelligence
  • Digital domestic policy in Europe
  • Deepfake bans and fundamental rights protection
  • Industrial competitiveness

Clarus Lead

The failure of negotiations reveals a fundamental conflict between industrial policy interests and fundamental rights protection in Europe. Germany and the EPP group are using their blocking power to protect national champions like Siemens and Bosch from additional regulations – with the risk that no protection mechanism against sexualized deepfakes will be enacted. The legal uncertainty simultaneously jeopardizes the goal of establishing European standards as a global benchmark for trustworthy AI.

Detailed Summary

Core Dispute over Sector Exit and Double Regulation

The EPP group under Merkel's successor is demanding that industrial AI applications – particularly in medical technology and mechanical engineering – be exempted from the AI Act. Justification: These products are already subject to strict sector-specific laws; additional AI requirements merely create bureaucracy. The sector exit is intended to protect German heavyweights from competitive disadvantage.

Opponents – Greens, Social Democrats, and several EU countries – warn of a patchwork of different standards. The TÜV Association emphasizes: The horizontal approach of the AI Act (unified high-risk regulation across sectors) is the core of the regulation. A retreat into sector-specific individual rules creates a long regulatory vacuum and weakens Europe's industrial standing rather than strengthening it. Additionally, such a move jeopardizes the chance to establish European standards as a global norm.

Societal Blockade: Deepfakes and Fundamental Rights

Particularly critical: The blockade also delays the planned ban on sexualized deepfakes and AI-powered undressing apps. AlgorithmWatch criticizes that the omnibus was going to be rushed through anyway without adequately examining fundamental rights consequences. The organization demands that a deepfake ban should not fail on technical definitions of "intimate body parts" – the key must be explicit consent of affected persons. The delay provides room for improvements but also increases the risk that no protection mechanism will be enacted at all.

Key Statements

  • German EPP group blocks trilogue negotiations on AI reform through demand for sector exit for industrial applications
  • Opposition from Greens, Social Democrats, and TÜV Association: Horizontal AI regulation is the core of the AI Act; patchwork weakens Europe
  • Blockade simultaneously prevents ban on sexualized deepfakes and delays fundamental rights protection
  • Commission and business associations push for quick return to negotiating table; timeframe unclear

Critical Questions

  1. Evidence/Source Validity: The claim that sector-specific laws are sufficient for high-risk AI scenarios – are examples (medical technology, mechanical engineering) supported by available data, or are they assumptions?

  2. Conflicts of Interest: To what extent do Siemens, Bosch, and other German corporations specifically benefit from a sector exit? Are analyses available that quantify the economic advantage versus regulatory costs?

  3. Causality/Alternatives: Is the sector exit truly necessary to ensure competitiveness, or are there compromise options (e.g., differentiated requirements instead of complete exemption)?

  4. Feasibility/Side Effects: How would a fragmented regulatory framework (sector-specific vs. horizontal) work in practice? Do multiple regulations actually create compliance costs, or savings?

  5. Fundamental Rights Protection: To what extent are technical definitions ("intimate body parts") sufficient for a deepfake ban, and who bears the burden of interpretation?

  6. Political Feasibility: What scenarios exist for resuming negotiations? Can the EPP position be enforced without compromise?

  7. Timing Risk: The original deadline was August 2024 – how do further delays affect practical implementation?

  8. European Standard: Does Europe actually lose global standard-setting power if it regulates internally in a fragmented manner?

Bibliography

Primary Source: AI Act: Europeans Cannot Agree on Digital Omnibus – heise online, Stefan Krempl

Referenced Organizations & Actors:

  • EU Parliament (Trilogue Negotiation)
  • European Commission (Trilogue Negotiation)
  • EPP Group (Blocking Position)
  • TÜV Association (Criticism)
  • AlgorithmWatch (Fundamental Rights Protection)
  • DOT Europe (Platform Lobby)

Verification Status: ✓ 2025

This text was created with the support of an AI model. Editorial Responsibility: clarus.news | Fact-Checking: 2025