Trump Signs AI Executive Order, Perplexity Builds New Load-Shifting Tech, EU Data Center Plan Struggles — AI News Roundup for June 3

Trump Signs Executive Order Strengthening AI Oversight — Scaled Back After Industry Opposition

U.S. President Trump signed an executive order on June 2 strengthening government oversight of advanced AI models. An early draft had considered mandatory pre-release reviews, but after strong opposition from industry groups, the final version was scaled back to call for **voluntary government reviews** ahead of release.

The story was major breaking news covered across multiple outlets, with the New York Times, Wall Street Journal, and TechCrunch each reporting from their own angles. The government's push for "early access" to cutting-edge AI models has supporters on national security grounds, but concerns about protecting trade secrets and stifling innovation remain persistent.

Perplexity Develops New Platform That Splits AI Processing Between PCs and Servers

Bloomberg reports that Perplexity AI has developed a platform that dynamically distributes AI tasks between PCs and cloud servers to cope with exploding demand for AI computing.

The software the company built is said to act as a kind of "air traffic control" for AI tasks. When demand spikes, it offloads some processing to edge devices to ease the load on cloud infrastructure. With soaring AI inference costs becoming an industry-wide challenge, the approach is drawing attention as a practical solution.

EU's €20 Billion AI Data Center Plan Hits Turbulence

Bloomberg reports that the EU's planned €20 billion (about ¥2.3 trillion) mega AI data center initiative is not progressing smoothly, beset by delays and funding problems.

The plan to build five massive data centers — so-called "gigafactories" — is reportedly seeing some partner companies begin to walk away. The initiative is a cornerstone of Europe's efforts to boost its AI competitiveness, and delays could ripple across the EU's entire AI strategy.

Google's Gemini Spark AI Agent Draws Buzz as "Amazing, and Creepy"

A reporter at The Verge put Google's AI agent, Gemini Spark, to the test and judged it "impressive and unsettling at the same time."

In a travel-planning demo, the agent picked up on the user's intent through natural conversation and carried out tasks autonomously. As AI agents grow more capable, some point to a psychological discomfort with their autonomy — raising, once again, the question of how to balance technological progress with acceptance.

AI-Generated Paper Problems Surface at NeurIPS 2026

Submissions of AI-generated papers have become an issue in the position paper track at NeurIPS 2026, a top machine learning conference.

NeurIPS organizers have addressed the matter on their official blog and say they are considering how to respond. As AI-assisted paper writing becomes commonplace, the debate centers on where to draw the line — and how to protect the integrity of academic research.