Discover the Latest AI News📰

Tencent has unveiled QBotClaw, China's first AI browser that lets users freely connect to major large language models. This innovative tool transforms browsers from simple web portals into intelligent assistants capable of complex tasks. With cross-platform functionality and visual recognition features, QBotClaw represents Tencent's ambitious push into the evolving AI agent landscape while making advanced technology more accessible.

Google's AI-powered search overviews now boast a 90% accuracy rate, but that still leaves millions of potential errors every hour. A recent evaluation found that while accuracy improved from Gemini 2 to Gemini 3, inconsistencies with source material actually increased. Journalists have demonstrated how easily these systems can be tricked by fake content, raising concerns about reliability in an era of rampant misinformation.

Microsoft's Bing team has unveiled Harrier, a groundbreaking multilingual embedding model now available as open source. Supporting over 100 languages, this AI powerhouse leverages GPT-5 synthetic data and boasts impressive 32,000-token context windows. With versions ranging from 60 million to 2.7 billion parameters, Harrier promises to revolutionize search engines and AI agent services while remaining accessible across hardware capabilities.

Spanish startup Xoople has secured $130 million in fresh funding, pushing its total capital to $225 million. The company is building a satellite network that will feed real-time Earth data to AI systems through its EarthAI platform. Partnering with defense contractor L3Harris, Xoople aims to revolutionize everything from agriculture forecasts to urban planning by creating what it calls an 'Earth Record System' - a digital twin of our planet constantly updated by orbiting sensors.

Google is bringing its Gemini AI to Gmail while making bold privacy promises. The tech giant vows emails will never train its AI models, with data processed in isolated environments that self-destruct after use. As concerns grow about AI handling sensitive information, Google positions this as a safer alternative to competitors' offerings. The integration offers practical email assistance features while attempting to address the industry's trust deficit around AI and privacy.

GitHub has introduced an innovative AI feature called Rubber Duck that acts like a coding partner, reviewing your work through different AI models to catch errors you might miss. Early tests show it can improve code accuracy by nearly 75%, especially helpful for complex programming tasks. Developers can choose between automatic reviews or request feedback whenever they need a second opinion on their code.

Shanghai Wanmeng Technology has launched its 'Magic Cube AI Quality Inspection VOC' on Taobao's service marketplace. This innovative tool combines automated quality checks with customer feedback analysis, helping merchants improve service while uncovering valuable business insights. Moving beyond simple compliance monitoring, it promises to transform customer service into a strategic growth driver for e-commerce businesses.

At China's Restaurant Chain Summit, Meituan's Li Shubin revealed why digitizing basic business information - not flashy AI tools - is the real first course for restaurants. As diners' requests grow more complex ('Find me a pet-friendly French bistro with sunset views under ¥200'), AI needs accurate digital menus, inventory and ambiance tags to make smart matches. Meituan's investing billions to build this 'local life information infrastructure' so every eatery can eventually have an AI sous-chef handling customer quirks.

A Chengdu programmer has developed an innovative traffic violation reporting app using AI vision technology. Pu Haiyang's creation compresses manual reporting processes from minutes to seconds, with 90% accuracy in identifying infractions. The app features auto-zooming target tracking and direct integration with traffic authorities' systems while maintaining user privacy through local data storage. Expected to launch within three months, this tool could revolutionize how citizens participate in traffic management.

Google has rolled out a crucial update to its Gemini chatbot, transforming it into a faster pathway to mental health support for users in crisis. The move comes after troubling incidents involving AI interactions, prompting Google to simplify access to suicide prevention resources with a one-click interface. Alongside technical improvements, the company is committing $30 million to bolster global crisis hotlines. While this represents progress, questions remain about AI's ability to truly safeguard vulnerable users.

Elon Musk has launched a legal battle against OpenAI, filing a lawsuit that accuses CEO Sam Altman and President Greg Brockman of fraud. The Tesla billionaire claims OpenAI strayed from its non-profit roots after he donated $38 million, and now wants the court to remove its leadership and restore its original mission. This latest clash marks another chapter in Musk's rocky relationship with the AI company he co-founded but left in 2018.

The UK's ambitious National Data Library project, designed to fuel AI innovation, is hitting roadblocks due to poor-quality public datasets. Researchers found inconsistent labeling, outdated information, and accessibility problems that could force AI systems to rely on less reliable sources. While the government has pledged £100 million for the initiative, experts warn that without significant improvements in data quality and standardization, this opportunity might be wasted.

Alibaba Group has announced a major restructuring of its AI divisions, appointing renowned computer scientist Fei-Fei Li as Alibaba Cloud's CTO while elevating its Tongyi Lab to a full business unit. The moves signal Alibaba's aggressive push into artificial intelligence, combining top global talent with strategic organizational changes. These developments come as Alibaba's Qwen models gain international recognition, positioning the company for what it calls 'the second half of the AI era.'

China has unveiled a groundbreaking carbon accounting system that promises to transform how we measure global emissions. Developed by the Shanghai Institute of Advanced Technology, this innovative model combines massive data processing with advanced algorithms to provide more accurate assessments of carbon footprints across industries and borders. Early results suggest it may significantly alter our understanding of national emissions responsibilities, potentially giving China stronger footing in international climate negotiations.

Chinese SaaS provider Weimob has launched a specialized AI tool for retailers that simplifies complex operations through natural language commands. The 'Weimob Admin Skills' platform integrates with the OpenClaw ecosystem, transforming traditional menu-driven retail management into conversational AI-powered workflows. Merchants can now handle inventory, sales tracking, and staff performance with simple voice or text prompts, marking a shift from manual configuration to intelligent automation in retail operations.

Anthropic has unveiled Mythos, its most advanced AI model yet, currently available only to select security partners. This powerful tool has already identified thousands of previously unknown vulnerabilities in software code, some dating back decades. While demonstrating remarkable potential for cybersecurity defense, concerns linger about its potential misuse. The company is navigating complex discussions with US officials while maintaining strict controls over access to this groundbreaking technology.

A Chinese cybersecurity team has uncovered critical vulnerabilities in the OpenClaw AI agent, including one high-risk flaw that could allow attackers to take control of user devices. The findings highlight growing security challenges as AI agents become more integrated into daily workflows. Unlike traditional scanning tools, the detection system used mimics human security experts' intuition, marking a shift toward more intelligent vulnerability discovery methods.

A mysterious new AI video model called Happy Horse has unexpectedly topped the AI Video Arena leaderboard, outperforming the previously dominant Seedance2.0. The model excels in video consistency, detail accuracy, and motion naturalness while offering native audio generation. While its developers remain unknown, industry experts suspect Asian origins. This surprise contender signals accelerating competition in AI video generation technology.

Google Maps is rolling out a clever new feature that uses AI to automatically generate captions for your shared photos and videos. Powered by Gemini technology, this tool analyzes your images and suggests descriptive text, which you can edit or approve with a tap. Currently available for iOS users in the U.S., the feature aims to make sharing location experiences easier while maintaining personal touches. Google plans to expand it globally and to Android soon, alongside other user-friendly updates to their contribution system.

Anthropic has unveiled Mythos, its most powerful AI model yet, specializing in uncovering hidden software vulnerabilities. This digital detective can spot flaws even in decades-old code, outperforming human experts. But its capabilities come with risks - the same tech that could protect systems might also be weaponized. Currently limited to select tech giants and government partners, Mythos is sparking debates about AI ethics and security in an increasingly vulnerable digital world.