AI-News-posts

More than 1,000 Google DeepMind staff tied to London have backed a union drive centered on military AI ethics, policy changes at Google, and bargaining rights. Organizers are seeking CWU and Unite recognition and may pursue legal steps if talks stall.

Greg Brockman’s defense of an implied multibillion-dollar OpenAI stake spotlights a governance model that separates control from profit. Here’s how the nonprofit, capped‑profit, and Public Benefit Corporation setup works, and why it matters for employees, partners, and investors.

Reports indicate Anthropic’s high-capability cybersecurity model is being tested or used by the NSA while access remains tightly restricted. A vendor access incident underscores supply-chain risk and fuels policy debates about who should get powerful offensive security AI.

Investigations point to a coordinated pro‑AI political network powered by super PACs and 501(c)(4) nonprofits, bankrolled by major tech and crypto donors. The money and messaging infrastructure could amplify influencer-style campaigns while masking who pays.

Court filings and reporting shed light on how Shivon Zilis maintained a confidential advisory channel with Elon Musk while serving on OpenAI’s board. Here’s what the record shows and why it matters for AI governance.

NVIDIA’s OpenClaw ecosystem, built on the open source NemoClaw stack and Nemotron models, points to a practical path for secure, cost-aware AI agents that run locally and tap frontier cloud models only when needed. Early case studies show sizable gains in verification, security operations, and commerce.