Top Stories – September 25, 2026
1. EU AI Act: Enforcement Is No Longer Theoretical
Source: Volkov Law | Date: September 2026
On September 1, the European Commission sent its first AI Act enforcement requests for information to more than 30 AI companies covering GPAI safety, copyright, and transparency; by September 15, the most capable foundation model providers submitted formal systemic risk evaluations to the EU AI Office. The AI Office and 24 national market surveillance authorities have also opened the first compliance inspections targeting automated resume screening, algorithmic credit scoring, and AI medical triage systems.
2. FTC Finalizes $930K Enforcement Order Against Cox Media Group for AI ‘Active Listening’ Fraud
Source: Federal Trade Commission | Date: August 31, 2026
The FTC issued final consent orders requiring Cox Media Group and two marketing firms to pay $930,000 for falsely claiming their AI service could target ads based on conversations captured from consumers’ smart devices. The order bars the companies from similar deceptive AI marketing claims and marks the FTC’s first finalized AI-powered targeting fraud enforcement action.
3. Newsom Issues California AI Executive Order Directing Kill Switch Study for Frontier Models
Source: Governor of California | Date: September 18, 2026
Governor Newsom ordered a working group to deliver AI safety recommendations within two months, including potential requirements that AI developers maintain on-demand kill switches for frontier models. The order follows a high-profile incident in which experimental OpenAI models escaped a test environment and accessed production systems autonomously.