Category Research

Coverage of the latest AI research, studies, and technological breakthroughs.

White House Accuses Moonshot AI of Copying Anthropic as OpenAI Security Test Goes Wrong

AI model theft claims against Moonshot AI, OpenAI containment issues, and rising token costs are reshaping the U.S.-China race.

Anthropic launches Opus 5 with stronger benchmarks, fewer restrictions and new fallback tools for safer, smoother AI use.

AI security fears surged as Kimi K3 spooked Wall Street and an unreleased OpenAI model was linked to a Hugging Face breach.

Trump’s AI science funding push sends $5 billion to Genesis Mission projects, but researchers warn it could weaken U.S. science.

Runway launches an AI model router for generative media, betting developers want one layer to choose the best image, video or audio model.

Kimi K3 is at the center of claims it copied Anthropic’s Fable, but researchers say distillation alone does not explain its rapid rise.

Chinese open-weight AI models are nearing U.S. frontier systems, rattling Silicon Valley, Washington and the business case for closed AI.

Arcee says open-weight models from China are not inherently dangerous and should be judged through standard enterprise security checks.

OpenAI says its AI cybersecurity test models escaped a sandbox and reached Hugging Face, raising new concerns about agent containment.

OpenAI says its own models caused an AI breach at Hugging Face during a cyber test, exposing new risks in frontier model evaluation.