The "Grok" Amendment and the 2026 AI Omnibus: A Mid-March Update

The Grok Scandal and the New Prohibited Practices

Yesterday, March 11, EU lawmakers reached a pivotal political agreement on a package of amendments to the EU AI Act. Driven by the "Grok affair"—a controversy involving the ease with which users generated non-consensual sexual deepfakes—the Act now includes an explicit prohibition on non-consensual intimate AI images.

This is a major shift. Originally, the Omnibus was intended to be an administrative "clean-up," but the Grok scandal forced regulators to close a loophole that had allowed realistic synthetic sexual content to evade the "high-risk" classification.

Industrial Relief: The AI Omnibus Pivot

While social media platforms face tighter rules, industrial manufacturers received a "boost for competitiveness." The new deal includes eased compliance for AI systems embedded in sector-regulated products, such as:

• Medical Devices

• Industrial Machinery

• Automotive Systems

These sectors will now largely use their existing sectoral regulatory frameworks to demonstrate AI compliance, avoiding the "double or triple layer" of regulation that trade groups like DIGITALEUROPE warned could cost SMEs up to €319,000 in initial filing fees.

ISO 42001: The Leading Framework for 2026

In the private sector, the momentum for ISO/IEC 42001 is accelerating. This week, MIND became the first data security company to achieve certification, and global law firm K&L Gates followed suit. These certifications demonstrate that ISO 42001 has become the bridge for organizations that need to satisfy both the EU AI Act and the U.S. Executive Order on Safe AI simultaneously.

← Back to resources