2 August 2026: What the EU AI Act Changed, and What Moved to December 2027
Article 50 transparency obligations applied from 2 August 2026. The high-risk obligations did not: the Digital Omnibus moved them to 2 December 2027.
INSIGHTS
AI Governance
Insights on AI compliance, regulation, and auditability.
More Articles
EU AI Act Article 12: What Infrastructure-Level Compliance Actually Requires
Documentation platforms generate reports. Article 12 requires infrastructure that captures events automatically. The distinction matters when auditors arrive.
The Multi-Model Governance Challenge
Your enterprise uses Claude, GPT-4, and Gemini. But whose governance tools are you using? And what are they missing?
Federal Preemption Meets State Innovation: The Regulatory Patchwork That Makes Infrastructure-Level Governance Essential
Trump's EO 14365 creates federal preemption while NY's RAISE Act demands 72-hour incident reporting. Organizations need infrastructure-level governance that works across any jurisdiction.
100 Hallucinated Citations: The Academic Trust Crisis Mirroring Enterprise Risk
GPTZero found 53 NeurIPS papers with 100+ fabricated citations. With 17% of peer reviews AI-written, verification is failing at the world's top AI conference.
Deloitte's A$440,000 Hallucination: When AI-Generated Reports Cite Non-Existent Research
A Deloitte government report contained AI-generated fabricated academic sources and invented quotations. The second such incident on two continents reveals a verification infrastructure gap.
California Becomes First State to Regulate Frontier AI: What SB 53 Means for Every AI Team in America
Governor Newsom signed SB 53 into law, making California the first state to regulate frontier AI developers with transparency, safety testing, and incident reporting requirements.
Want to discuss AI governance?
Talk to us about compliance strategy and technical questions.
Talk to us