US AI Brief — May 14, 2026
Top Stories
1. U.S. and China Establish AI Safety Protocol as Bessent Declares American Lead
CNBC · May 13, 2026
Treasury Secretary Scott Bessent announced from the Beijing presidential summit that the U.S. and China will establish a protocol for AI safety best practices, specifically aimed at preventing non-state actors from accessing advanced models. Bessent framed the cooperation as viable because “we are in the lead” technologically, adding that the dialogue would not occur if China were ahead. He also signaled upcoming “step-function jumps” in LLMs from Google Gemini and OpenAI, while noting ongoing “back and forth” over Nvidia chip export restrictions.
Why It Matters: This marks the first major U.S.-China AI governance framework acknowledging superpower competition while establishing critical safeguards. Bessent’s confidence underscores Washington’s strategic posture as export controls remain contested.
URL: U.S. can hold AI talks with China because ‘we are in the lead,’ Bessent tells CNBC
2. Turf War Erupts Over Who Polices Frontier AI Models
Risky Biz · May 14, 2026
The Trump administration is divided over whether the Commerce Department or intelligence community should lead pre-release assessment of frontier AI models. Commerce’s CAISI (formerly the U.S. AI Safety Institute) had announced voluntary testing agreements with Google, Microsoft, and xAI but was forced to remove the website amid White House “sensitivity.” The National Cyber Director has proposed placing evaluation within the Office of the Director of National Intelligence, while Commerce defends its existing infrastructure. Sources describe the conflict as a “knife fight” between agencies.
Why It Matters: The outcome determines whether AI companies face pre-release government review — a meaningful escalation from current light-touch oversight. The intelligence community brings cybersecurity expertise, but AI’s implications across the entire economy may require a broader approach.
URL: Srsly Risky Biz: The AI Regulation Knife Fight
3. Google Brings Agentic AI and “Vibe-Coded” Widgets to Android
TechCrunch · May 12, 2026
At its Android Show: I/O Edition, Google announced Gemini Intelligence features enabling cross-app task completion, web browsing, form filling, and natural-language widget creation. Users can press the power button to execute multi-step processes — copying grocery lists from notes apps and adding items to shopping carts — with Gemini waiting for final confirmation before checkout. The “vibe-coding” feature allows users to describe widgets in plain text, such as “Suggest three high-protein meal prep recipes every week.” These features will first arrive on Samsung Galaxy and Google Pixel devices this summer.
Why It Matters: Agentic AI on mobile devices represents the next battleground for consumer AI adoption, with Google leapfrogging assistant capabilities. Enterprise mobile strategies must account for AI-native user behaviors emerging from these capabilities.
URL: Google brings agentic AI and vibe-coded widgets to Android
