← All Issues
Wednesday, May 7, 2026

AI Daily

Your Automated Intelligence Briefing

Geopolitics

Manufacturing the China Panic: Silicon Valley's $125 Million Fear Campaign

OpenAI and Palantir executives have quietly bankrolled a campaign paying TikTok influencers up to $5,000 per video to stoke fears about Chinese AI. The timing is not random: the campaign launched as local opposition blocked $64 billion in US data center projects, and as Stanford's AI Index found the US-China performance gap had collapsed to 2.7%.

Read full story →
Policy

Before the Launch Button: Why Governments Are Now Testing AI Models First

Google, Microsoft and xAI have joined OpenAI and Anthropic in letting the US government evaluate their AI models before public release. The shift was accelerated by Anthropic's Mythos model finding 300 vulnerabilities in Firefox alone, a number fifteen times higher than prior models. The question nobody has answered: what happens when the evaluation finds something genuinely alarming?

Read full story →
Privacy

Chrome's Uninvited Passenger: The 4GB AI Model Nobody Asked to Install

Google Chrome has been silently downloading Gemini Nano, a 4GB AI model, onto user machines without notification or consent. The model re-downloads itself if manually deleted. Security researcher Alexander Hanff found similar silent expansion in Anthropic's Claude Desktop, and the contrast with Apple's explicit AI model choice in iOS 27 makes the industry defaults look increasingly hard to defend.

Read full story →
Opinion — Peter Harrison
Opinion

Team USA and the Displacement It Does Not Mention

Silicon Valley's new patriotism campaign tells workers that Chinese AI is the threat. It says nothing about the automation those same companies are deploying in American workplaces. The geography of the model that takes your job does not change what it takes.

Read opinion →