hotai.news
中文
AI Briefings / Date Archive

2026-09-07
AI Briefings Daily Report.

An archive of 20 AI news briefs collected on 2026-09-07. Check the original sources for details and updates.

This content still needs to be cross-checked against the original publication page.

20 curated selectionsView other dates ↗
01News

Alibaba Releases Qwen-Drive 1.0, a Single Model Handling Both Driving and Cabin Assistant Functions

Alibaba's research arm has released Qwen-Drive 1.0, an AI model that handles environmental perception, traffic Q&A, and route planning in one system. The researchers show that text-image models don't automatically understand three-dimensional space. Spatial awareness has to be trained on purpose. The goal is a single model that runs both the cockpit and the driving system.

Source links can be verified on the original siteRead original ↗
02News

OpenAI discloses researcher AI coding agent usage; top 10% of users consume Tokens worth over 7000 US dollars per day

OpenAI for the first time discloses the scale and cost of its researchers using AI coding agents; high-frequency users consume over 7000 US dollars worth of Tokens daily, and ordinary researchers average over 600 US dollars daily. Since the beginning of the year, post volume in the internal tech support channel has dropped by more than half, as AI profoundly changes R&D work methods.

Source links can be verified on the original siteRead original ↗
05News

OpenAI Chief Scientist Jakub Pachocki publishes article urging slower AI development, warning agents may run out of control

OpenAI Chief Scientist Jakub Pachocki published a warning that AI agents may deceive humans, bypass monitoring, and even accelerate their own evolution. He called for mandatory safety thresholds and mentioned Anthropic's runaway case. #AIsafety# sparked heated discussion. OpenAI CEO Sam Altman reposted and liked.

Source links can be verified on the original siteRead original ↗
06News

WeChat internal testing"Xiaowei AI Social"Feature that lets AI chat with your friend's AI on your behalf

A user's Xiaowei can directly find a friend's 'Xiaowei' to chat; just tell one's own 'Xiaowei' the request, and it goes to the other party's 'Xiaowei' to communicate; after the other party confirms, the two 'Xiaowei' communicate first, then bring back the result; when user approval is needed, it reminds them to confirm.

Source links can be verified on the original siteRead original ↗
10News

Alibaba Qwen Office Launches Industry's First Multi-Person Workbench, Supporting Online Collaboration for Hundreds of Users

Alibaba Cloud Qianwen Office introduces the industry's first "multi-person workbench," supporting simultaneous online collaboration for up to 100 users. It applies to scenarios such as event organization, home-school coordination, and enterprise collaboration, reducing SaaS procurement and custom development costs.

Source links can be verified on the original siteRead original ↗
12News

OpenAI releases internal research acceleration measurement data: research organization agent workload has reached 3.1 times a human workday

For AGI to benefit all of humanity, we believe it must be democratically governed. This can only happen through an informed public debate about the capabilities, risks and safeguards of highly capable AI systems. People everywhere need to understand the likely future trajectory of frontier AI, so they can have a meaningful voice in how it develops.

Source links can be verified on the original siteRead original ↗