Welcome to the "AI Daily" section! This is your guide to exploring the world of artificial intelligence every day. Every day, we present you with the latest content in the AI field, focusing on developers to help you understand technology trends and learn about innovative AI product applications.
Fresh AI products Click to learn more:https://app.aibase.com/en
1、Shanghai's first AI prompt copyright case ruled: prompts lack originality
This article discusses the first-instance judgment made by the Huangpu District People's Court in Shanghai regarding a copyright infringement case involving AI prompts. The court ruled that the prompts lacked originality, so the plaintiff does not hold the copyright.
【AiBase Summary:】
⚖️ The court ruled that prompts lack originality, and the plaintiff does not have copyright.
📝 The defendant argued that the prompts are not protected by copyright and fall within the scope of fair use.
🌟 The plaintiff is an art creation company that sued the defendant for copyright infringement due to works generated by prompts.
2、New Thinking Model "Moonshot Kimi K2 Thinking" Released, AI Boundaries Pushed Further!
Moonshot released its latest thinking model, Kimi K2 Thinking, which achieved significant improvements in intelligent agents and reasoning capabilities. It can autonomously perform multi-round tool calls and deep thinking without user intervention, marking a step forward in AI technology toward higher levels of self-learning and adaptability.
【AiBase Summary:】
🧠 Kimi K2 Thinking can think and use tools simultaneously without user intervention.
🚀 It improves the autonomy and efficiency of AI when handling complex tasks.
💡 It has attracted industry attention and is expected to bring new ideas in fields such as healthcare, finance, and education.
3、A New King in Chinese Image Editing! UniWorld-V2 Released, Select and Edit with a Box, Accurate Rendering of Chinese Fonts, Performance Outperforms GPT-Image and Gemini
The article introduces the UniWorld-V2 image editing model and its technical breakthroughs based on the UniWorld-R1 framework, showcasing its excellent performance in understanding Chinese instructions and image editing capabilities.
【AiBase Summary:】
🧠 The UniWorld-V2 model significantly improves the accuracy and flexibility of image editing through an innovative visual reinforcement learning framework called UniWorld-R1.
🖌️ The model can accurately understand Chinese instructions and achieve complex Chinese font rendering and lighting effects.
📈 In test benchmarks GEdit-Bench and ImgEdit, UniWorld-V2 leads other well-known models like GPT-Image-1 and Gemini2.0 with high scores.
Details link: https://arxiv.org/abs/2510.16888
4、Google Releases AI File Detection Tool Magika 1.0, Major Upgrade Fully Uses Rust Language
Google's Magika 1.0 is the latest stable version of its AI-based file type detection system, with significant improvements in performance and security. Its core engine has been migrated to the Rust language and supports over 200 file formats, including new types for data science and programming languages. Additionally, Magika 1.0 simplifies the developer integration process and encourages community participation in optimization.
【AiBase Summary:】
🌟 Magika 1.0 is restructured using the Rust language, improving performance and security.
📂 Supports more than 200 file formats, adding new types for data science and programming languages.
⚙️ Simplifies the developer integration process and encourages community participation in project optimization.
5、Sora Android Version Hits Explosive Downloads on Launch Day, Reaching 470,000 Times!
Sora's Android version reached 470,000 downloads on its launch day, performing exceptionally well, while its iOS version had lower downloads due to regional restrictions and an invitation-only system. Sora provides personalized video generation services through AI technology and faces competition from Meta AI in the short video field.
【AiBase Summary:】
📱 Sora Android version reached 470,000 downloads on the first day, showing strong performance.
🌐 iOS version is limited to the United States and Canada, and uses an invitation-only system to restrict user access.
🎥 Sora offers AI-generated video features, supporting animated video creation and interactive browsing.
6、The World's First AI Ocean Large Model "Kanhai" Launched! 10-Day Ocean Forecast Accurate to 600 Meters Deep
The article introduces China's independently developed global first end-to-end artificial intelligence ocean large model "Kanhai," which can reconstruct current sea conditions in real time and accurately simulate changes in temperature, salinity, density, and flow fields from the sea surface to a depth of 600 meters over the next 10 days, providing unprecedented support for ocean research, disaster prevention, and ecological management.
【AiBase Summary:】
🌊 First implementation of a full-loop closed system from satellite remote sensing observations to three-dimensional ocean environment prediction.
🧠 Utilizes an end-to-end intelligent inference architecture, skipping traditional modeling steps, improving prediction efficiency and accuracy.
🌍 Provides a high-precision "digital ocean mirror" for ocean research, disaster prevention, and ecological management.
7、Iron Man and Robot Punch Become Reality! Yushu Technology's "Embodied Avatar" Makes Human Avatar Robots Awaken Instantly
The article introduces Yushu Technology's Embodied Avatar system, which enables remote control of robots through high-precision motion capture and 5G technology, featuring low latency and high precision, with broad application prospects in multiple fields.
【AiBase Summary:】
🤖 Full-link low-latency motion mapping technology enables remote control of robots.
🧠 Self-developed UnifoLM model enhances robot's autonomous adaptability.
🚀 Applications cover industrial, medical, educational, and entertainment fields.
8、Google Launches Gemini API File Search Tool: Simplify Private RAG Integration, Developers Don't Need to Build Their Own Vector Databases
Google launched the file search tool for Gemini API, a fully managed RAG system that seamlessly converts private files into knowledge bases. This tool simplifies the data processing workflow, allowing developers to focus on application logic rather than maintaining underlying infrastructure.
【AiBase Summary:】
🌍 Supports various file formats, including PDF, DOCX, TXT, and programming language files.
🔍 Automatically processes file uploads, indexing, and retrieval, supporting semantic search.
💰 Innovative billing model, first indexing cost as low as $0.15 per million word tokens.
