AI News

Google has announced Gemini 4 Argon with a one-million-token output limit and introductory API pricing. Selected cyber defenders get access first; the broader rollout has no announced date.

Anthropic warned that AI could pose existential risks to humanity in IPO documents, raising questions about safety disclosures and investor scrutiny.

Google Research releases RRSI, an Apache 2.0 framework that evolves AI agent harnesses while using regularization to limit benchmark overfitting.

AI data centers are replacing farmland and undeveloped land, driving power and water concerns and fueling opposition across U.S. communities.

Google Research combines four AI agent frameworks for coherent minutes-long video, using visual memory, planning, critique, and prompt refinement.

Nvidia’s Open Agent Safety Platform adds permission controls and monitoring to help companies contain AI agents and prevent sandbox breakouts.

Bill Gates says AI companies cannot regulate themselves alone and urges lawmakers and law enforcement to require safeguards and monitoring.

Sarvam’s Saaras V4 supports 22 Indian languages, global English, five output modes and API transcription at ₹30 per hour.

Kenjiro Tsuda is suing TikTok over videos allegedly using an AI clone of his voice, testing Japan’s protections for performers and vocal identity.

A Fox News discussion examines the U.S.-China artificial intelligence race, Trump’s summit with Xi and growing concerns about AI oversight.