AI News Briefing — March 21, 2026

The Trump Administration released a federal AI legislative framework seeking to streamline regulations at the national level, aiming to preempt a patchwork of state-by-state AI governance. The framework could face resistance from states that already have their own AI regulations in place.

A Meta AI agent instructed an engineer to take actions that exposed a large amount of sensitive user and company data internally for two hours. The incident highlights growing risks as companies deploy autonomous AI agents in production engineering workflows.

Jeff Bezos and Blue Origin filed with the FCC for 'Project Sunrise,' a megaconstellation of up to 51,600 satellites for orbital data center services, arguing terrestrial AI data centers will struggle to scale. This follows SpaceX's similar proposal for up to 1 million satellites.

At Palantir's developer conference, the company doubled down on its vision of AI built for battlefield advantage, with business soaring and growing military/defense customer adoption. The event showcased deepening integration of AI into warfare and national security.

Atlassian laid off staff shortly after introducing AI 'teammates' agents into its workflows, with former employees noting the agents were helpful but insisting they couldn't replace human workers. The story highlights the emerging pattern of AI-driven workforce reductions at major tech companies.

Essex police paused live facial recognition after a study found black people were 'significantly more likely' to be identified compared to other ethnic groups. At least 13 UK police forces use the AI-enabled technology, raising broader concerns about systemic bias.

Uber invested $1.25 billion in Rivian, attracted by the electric carmaker's AI-driven autonomous and connected vehicle capabilities. The deal could help revive Rivian's finances after recent turbulence.