Microsoft AI Model Development Exposes 38TB of Sensitive Data

Translated data: Microsoft recently experienced a data breach on GitHub, resulting in the leakage of 38TB of private data, including passwords, keys, and internal messages, due to the misuse of shared access signature tokens (SAS) on the Azure platform. The improper use of SAS tokens and lack of monitoring exposed the data for years, highlighting the security challenges in AI model training. This incident underscores the need for stronger security measures and collaborative efforts in the AI development process that relies heavily on large-scale data.

GitHub Policy Reversal: Private Repository Code Will Also Be Used to Train AI

GitHub announced that starting April 24, 2026, it will use user interaction data to train AI models, causing controversy. This move has been jokingly referred to as a 'revert' operation, as GitHub had previously emphasized respect for user privacy. The new policy is opt-out by default, meaning private repository data will no longer be absolutely private, and even the free version of Copilot will be affected.

Tencent Cloud's Billing Strategy Undergoes Major Adjustment: Prices of Some AI Models Rise Significantly

The Tencent Cloud Intelligent Agent Development Platform will adjust its AI model billing strategy starting March 13, 2026. The key changes include ending the free trial of public test models and optimizing the pricing of self-developed Hunyuan series models. This move marks the maturity stage of Tencent Cloud's AI commercial ecosystem. Among them, three high-performance models, GLM5, MiniMax2.5, and Kimi2.5, will end their limited-time free public testing.

Tmall AI Institute Launches MAOSS Model: Using AI to Screen for Fatty Liver Disease, Doubles High-Risk Detection Rate

Alibaba Tmall AI Institute collaborated with multiple hospitals to develop the AI model MAOSS for screening fatty liver disease, and the research was published in Nature Communications. The model aims to address the challenge of early diagnosis of fatty liver disease, as traditional tests have insufficient sensitivity and high costs, while the prevalence of fatty liver disease is high and it can progress to severe liver diseases.

Microsoft Launches the Small Multimodal AI Model Phi-4: The Perfect Combination of Thinking and Perception!

Microsoft releases the open-source AI model Phi-4-Reasoning-Vision-15B, which has high-resolution visual perception and deep reasoning capabilities. It is the first small language model that achieves both 'clear vision' and 'deep thinking,' opening up new intelligent application scenarios for developers.

Google's New Version Gemini 3.1 Is Coming Soon, Are Developers Ready?

Google announced that the Gemini 3 Pro Preview version will be discontinued starting March 9. Developers are advised to migrate to the Gemini 3.1 Pro Preview as soon as possible. Users who use the -latest alias will automatically switch to the new version starting March 6, and the old version will be completely retired after March 9 to avoid service interruption. This adjustment has drawn attention from the developer community.