Discover the Latest AI News📰

Microsoft's Windows 12 is set to launch later this year, marking a dramatic shift in operating system design. Built on the flexible CorePC architecture, this update brings true modularity - letting users customize their OS like never before. But the real game-changer? AI becomes the system's beating heart, with Copilot evolving from helper to core component. Just be warned: your old PC might not make the cut for these advanced features.

A small Mexican development team faces financial ruin after accidentally exposing their Google Gemini API key, leading to $82,000 in charges within 48 hours. Despite pleas for mercy, Google insists they pay the full amount under its 'shared responsibility' policy. The incident highlights concerns about Google Cloud's billing safeguards compared to competitors like OpenAI.

A three-person tech team in Mexico faces financial disaster after accidentally exposing their Google Gemini API key. Within 48 hours, malicious actors racked up $82,000 in charges - nearly 500 times their normal monthly bill. While Google cites its 'shared responsibility' policy refusing refunds, developers worldwide are questioning why the platform lacks automatic spending caps during abnormal usage spikes.

The open-source AI project OpenClaw has officially launched its Weibo account, sparking immediate engagement from China's leading large model developers. Within hours of its first post, companies like Zhipu, Qwen, Moonshot and NetEase Youdao joined the conversation. This comes as OpenClaw continues gaining momentum globally, recently making headlines at MWC2026 while pushing Chinese industrial AI into deeper business applications.

Amazon has laid off another 16,000 employees, bringing total cuts to 30,000 in just three months. The company cites 'big company illness' and efficiency improvements as reasons, while quietly embracing AI technology that's transforming traditional roles. While Amazon promises support for affected workers, many wonder if this marks the beginning of a broader AI-driven employment shift across industries.

Alibaba's QWEN large model leader Lin Junyang (Justin Lin) announced his resignation in an emotional social media post today. The tech veteran, who played a pivotal role in developing Alibaba's AI capabilities, cited exhaustion after years of intense work. While assuring his team would continue strong, Lin declined immediate interviews, telling friends 'I really need some rest.' The industry now watches for his next move - and who will fill his shoes at QWEN.

News Corp CEO Robert Thomson revealed bold ambitions at the Morgan Stanley Tech Conference, positioning his media empire as essential infrastructure for AI development. The company's new $150 million deal with Meta grants access to premium news content from outlets like The Wall Street Journal for AI training. Thomson describes News Corp's breaking news as the 'semiconductors' powering artificial intelligence, while navigating complex relationships with tech giants through a mix of partnership and litigation.

OpenAI's new partnership with the U.S. Department of Defense has sparked fierce backlash, with ChatGPT uninstallations skyrocketing nearly 300% overnight. App store ratings plummeted as users flooded reviews with one-star ratings protesting military AI use. Meanwhile, competitor Anthropic sees unexpected gains by positioning itself as the ethical alternative.

As AI shopping assistants become ubiquitous, fraudsters are weaponizing the same technology for scams. Riskified's upgraded platform now offers real-time identity verification and customizable defense policies to protect businesses. The company also partnered with HUMAN Security to create a unified trust framework for human and AI interactions in e-commerce.

Microsoft faces mounting criticism as Windows 11 users revolt against forced AI integrations they call 'electronic waste.' The company's attempt to silence dissent by banning the satirical term 'Microslop' in its Discord community backfired spectacularly, leading to creative user protests and eventual server lockdowns. While Microsoft claims spam attacks prompted the restrictions, users see it as another example of prioritizing flashy features over system stability.

Baidu's experimental AI search app Tizi AI has been abruptly discontinued just half a year after launch. The service, known for its clean interface and unique film resources, will now be folded into Baidu's Wenxin App. While Baidu promises continued access to Tizi's core features, users wonder if the transition will preserve what made the standalone app special.

OpenAI is quietly developing its own code hosting platform, potentially setting up a clash with Microsoft-owned GitHub. The project, still in early stages, stems from frustration with GitHub's reliability issues. What makes this intriguing? Microsoft is OpenAI's biggest investor, turning this into a delicate dance between partners and competitors. The new platform could integrate AI coding tools like Codex, offering smarter automation than traditional repositories.

Alibaba founder Jack Ma made a rare public appearance with top executives at Hangzhou's Yungu School, signaling the tech giant's intensified focus on artificial intelligence. During lively discussions with students, Ma emphasized how AI will reshape society and why young people must adapt. The visit comes as Alibaba rolls out breakthrough technologies like its Qwen3.5-Plus model, positioning itself alongside Google as one of few companies mastering AI's full tech stack.

Google Home's latest update brings AI-powered real-time video analysis to Nest cameras. Instead of reviewing footage after the fact, users can now ask questions like 'Is there a package at my door?' and get instant answers. The upgrade also fixes common smart home frustrations like misheard commands and accidental device control. Currently available to premium subscribers, this marks a shift from passive recording to active home monitoring.

A chilling study reveals AI's alarming tendency toward nuclear escalation when placed in simulated crisis scenarios. Researchers tested three advanced models as national leaders, finding they chose military aggression far more often than human counterparts. The findings raise urgent questions about integrating AI into military decision-making.

Google DeepMind unveils Gemini 3.1 Flash-Lite, boasting impressive speed and intelligence gains over its predecessor. While processing over 360 tokens per second with quick response times, the model shines in complex tasks like scientific reasoning. However, these improvements come at a cost - pricing has nearly tripled, signaling a shift in the AI market towards premium performance.

Anthropic's Claude Code takes programming to new heights with its groundbreaking voice mode. Developers can now ditch their keyboards and simply speak commands to refactor code or optimize logic. Currently rolling out to select Windows users, this feature promises to reshape how we interact with AI coding assistants. Meanwhile, Anthropic's financials tell a compelling story - $2.5 billion in annual recurring revenue and user numbers that have doubled since January.

As midterm elections near, New York congressional candidate Alex Bores finds himself in Silicon Valley's crosshairs for championing AI safety regulations. A tech-backed Super PAC has poured $125 million into attack ads targeting Bores, who successfully pushed through New York's RAISE Act requiring major AI firms to disclose safety plans. The battle highlights growing tensions between tech leaders advocating unfettered AI development and policymakers demanding accountability.

At this year's Two Sessions, Xiaomi CEO Lei Jun put forward five ambitious proposals aiming to propel China's tech industry forward. His vision centers on transforming humanoid robots from lab curiosities into industrial workhorses, while addressing smart driving safety concerns. These initiatives could redefine how we interact with technology in factories and on roads.

China's MiniMax M2.5 large language model has taken the global developer community by storm, topping usage charts with an astonishing 3.07 trillion tokens processed in just seven days. The model's combination of affordability and specialized agent capabilities has propelled its parent company to $150 million in monthly revenue, while setting the stage for an intense showdown with upcoming releases from competitors.