OpenAI and Hugging Face disclosed a security incident during model evaluation, underscoring new risks in AI testing and the need for stronger safeguards.

The US says it may sanction Chinese open AI models over alleged IP theft, escalating AI policy from chip controls toward direct model scrutiny.

Google introduced Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and Flash Cyber, sharpening its low-cost AI lineup while Gemini 3.5 Pro stays in testing.

Box introduced new controls to govern AI agents accessing enterprise content, highlighting rising demand for safer AI automation in regulated workflows.

A thin Cybernews report says an OpenAI executive criticized Kimi and open-weight AI models, highlighting rivalry and messaging risks in AI competition.

Bloomberg reports OpenAI’s agents have reached 10 million users after the ChatGPT work launch, highlighting fast interest but limited public detail.

Microsoft and Mistral expanded their partnership to target regulated enterprises seeking frontier AI with more control, amid rising demand for open-model options.

Reports from The Guardian say China is narrowing the US AI lead, raising new questions for chip policy, model competition and enterprise bets.

Microsoft is reportedly funding Mistral’s European AI expansion in a multibillion-dollar deal, strengthening Azure’s role in regional AI infrastructure.

Alibaba has previewed Qwen3.8 and says the model ranks just behind Anthropic’s Claude Fable 5, signaling fresh pressure in enterprise AI.

New reporting says a researcher poisoned an open-weight AI model for under $100, underscoring supply-chain risks for enterprise AI teams.

The push to connect AI agents to enterprise apps is expanding automation potential while sharply increasing security, access, and oversight risks.

AISI says open-weight models now trail top closed AI cyber systems by just 4-7 months, shrinking defenders’ lead as costs fall sharply.

Ant, Tencent, Alibaba and Baidu are pushing AI work agents for enterprises, highlighting China’s shift from chatbots to workflow automation.

A Chinese AI model reportedly climbed to the top of rankings, underscoring rising pressure on OpenAI and Google as evidence remains thin.

This week's AI Week in Review cluster lacks enough verifiable source detail to support a factual reported story, highlighting evidence gaps readers should note.

French authorities warned that OpenAI, Google, and Anthropic dominate AI agents, highlighting lock-in risks for enterprise buyers and builders.

A brief Exponential View mention puts Kimi and Moonshot AI back on the radar, but limited sourcing makes the market significance hard to verify.

A report says OpenAI launched Codex Micro, a $230 RGB keypad for AI coding agents, signaling interest in dedicated hardware for developer workflows.

Chinese AI labs are gaining credibility, and CNBC says their progress is renewing attention on open-weight models for builders and enterprises.