Top Stories – August 9, 2026

1. White House Confirms EO 14409 Framework Delivered - Classified, Not Released Publicly

Source: Fortune | Date: August 4, 2026

The White House confirmed it met the August 1 EO 14409 deadline, finalizing a classified frontier AI model evaluation framework and briefing major AI labs – OpenAI, Anthropic, Meta, Nvidia, and Microsoft – on August 4. The benchmarks, model thresholds, and the framework document itself will not be released publicly, drawing criticism over the absence of any public accountability mechanism.


2. UK AI Regulation and Safety Bill Clears House of Commons

Source: Cubbbix | Date: August 2026

The UK AI Regulation and Safety Bill has passed the House of Commons, with Royal Assent expected by October 2026. The legislation would make mandatory safety data sharing for foundation model providers a legal requirement – a significant departure from the entirely voluntary framework in place since AISI’s launch.


3. EU AI Office Designates 8 Foundation Models for Monthly Systemic Risk Reviews

Source: Wilson Sonsini | Date: August 2026

The EU AI Office published a list of 8 foundation models that exceed the 10^25 FLOPs training threshold, now requiring their providers to submit monthly systemic risk evaluations under the AI Act’s newly active GPAI enforcement regime. Fines for systemic risk violations can reach EUR 15 million or 3% of global annual turnover.