📰 News(8857)

Chinese AI models are dominating global usage, with Tencent's Hunyuan model leading the pack. Recent data shows China's weekly token usage now surpasses the US by 1.8 times, marking three straight weeks of dominance. The competition heats up as price cuts and specialized models like the mysterious 'Owl Alpha' shake up the rankings.

Chinese edtech pioneer Squirrel AI has earned a spot on TIME's prestigious 2026 list of the world's most influential companies. Their adaptive learning platform, which has already helped 43 million students, uses AI to create customized education paths by identifying knowledge gaps in real-time. This recognition signals a shift in global education priorities from flashy tech concepts to practical learning outcomes.

Tencent has unveiled Ardot, an AI design tool that transforms natural language prompts into functional UI prototypes. Imagine describing an e-commerce homepage and getting a working design in minutes—that's the promise of this new platform. It bridges the gap between idea and implementation, allowing product managers, designers and developers to collaborate in real-time on the same canvas. While still in beta, Ardot could revolutionize how digital products are created by cutting through traditional workflow bottlenecks.

In a strategic move that could reshape financial services, Hang Seng Electronics has teamed up with Alibaba Cloud to accelerate AI adoption across the industry. The partnership combines Hang Seng's financial expertise with Alibaba's cutting-edge cloud technology, focusing on large language models and smart applications. Their first joint offering - an AI solution built on Qwen models - is already making waves in investment research and customer service sectors.

Chongqing recently hosted a landmark event as China Mobile IoT unveiled its vision internet AI ecosystem. Bringing together top universities and over 50 industry partners, the conference showcased ambitious plans to bridge AI algorithms with real-world applications. The company signed 38 strategic agreements, paving the way for smarter solutions in retail, safety supervision, and community governance.

Google has unveiled Gemini Intelligence, a game-changing AI suite for Android that can handle complex tasks like booking restaurants or comparing prices across apps. While the automation capabilities impress, they come with steep hardware requirements that could create a two-tier Android ecosystem. Only premium devices with 12GB RAM and flagship chips will support these features initially, raising questions about accessibility in the smartphone market.

Alibaba Cloud is gearing up to unveil a significant upgrade to its Tongyi Qianwen AI model at the upcoming summit on May 20. The company promises substantial improvements across multiple dimensions, signaling China's accelerating progress in artificial intelligence. This move comes as tech giants race to enhance their AI offerings amid intensifying competition in both cloud services and smart applications.

The academic preprint server arXiv is cracking down on low-quality AI-generated research papers with new enforcement measures. Authors caught submitting papers with obvious AI hallucinations or undisclosed AI assistance could face one-year submission bans. While not prohibiting AI use entirely, arXiv now requires authors to take full responsibility for all content - whether human or machine-written. The move comes as the platform transitions to independent nonprofit status to better address emerging challenges in academic publishing.

China's AI video generation technology is making waves globally, with creators praising tools from companies like ByteDance and Kuaishou for their superior quality and creative freedom compared to US counterparts. These platforms leverage vast short video libraries to train systems that excel at handling complex movements while maintaining stability. The technology is transforming advertising and e-commerce by enabling affordable, customized video content at scale.

Microsoft's AI leader Mustafa Suleyman has dropped a bombshell prediction: AI could replace most office jobs by mid-2027. In a recent interview, he forecast that professions like accounting, law, and marketing will achieve full automation as AI reaches human-level performance. While promising productivity gains, this vision raises urgent questions about workforce transitions and the future of knowledge work.

Hongguo Short Drama has refuted claims it eliminated base payments for AI-generated scripts, clarifying that some projects still qualify for guaranteed compensation. The controversy highlights growing tensions as AI reshapes content creation economics, with platforms balancing technological efficiency against creator incentives. While the company maintains certain script guarantees, its new performance-based models signal an industry shift toward data-driven compensation.

Security researchers have cracked Apple's latest M5 chip defenses in a record five days, using an AI-assisted data-only attack. The breakthrough exposes vulnerabilities in Apple's hardware-level memory protection, previously considered nearly impenetrable. This development signals a new era where AI accelerates both cybersecurity threats and defenses, forcing tech giants to rethink their security strategies.

When police in Hubei faced a diesel theft case with zero leads, they turned to an unlikely partner - ByteDance's AI tool Doubao. By analyzing wheelbase data from tire tracks, the system identified potential suspect vehicles, leading to a quick arrest. This breakthrough shows how AI is moving beyond office applications into real-world crime fighting, particularly where traditional methods fall short.

Linus Torvalds has raised serious concerns about how AI tools are overwhelming Linux kernel security teams with duplicate bug reports. The Linux creator warns that maintainers are drowning in repetitive submissions, forcing them to waste time confirming already-fixed issues. While not opposing AI use entirely, Torvalds urges developers to pair reports with actual code solutions rather than flooding channels with low-effort automated findings.

Sony clarifies how its Xperia 1 III's AI Camera Assistant actually works—it offers real-time shooting tips rather than editing photos. The feature analyzes lighting and subject characteristics to suggest exposure, color, and background adjustments. However, some users feel Sony's marketing overstated the AI's capabilities, particularly around 'best angle' recommendations. While the tech shows promise, early examples reveal room for improvement in color balance and natural-looking results.

Domino's faces a $100 million lawsuit from its largest U.S. franchisee after an AI-powered delivery system caused chaos. The 'Dragontail' platform, meant to streamline operations, instead led to cold pizzas and angry customers when drivers gamed the system. Riders began stacking orders and cherry-picking deliveries based on tips, turning what should have been a tech upgrade into an operational nightmare.

New research reveals a curious quirk in human psychology: we perceive AI as more confident than humans, even when giving identical answers. This 'confidence illusion' stems from our assumptions about technology's competence and the lack of emotional cues in AI responses. Scientists warn this blind trust could lead to poor decisions, prompting calls for clearer confidence indicators in future AI systems.

While tech giants pour billions into massive AI models, Sohu CEO Zhang Chaoyang charts a different course. The company is focusing on practical AI applications that boost efficiency while maintaining content integrity. This strategy reflects both financial realities and a commitment to responsible technology use in an industry often chasing the next shiny object.

Baidu's first-quarter results reveal explosive growth in its AI cloud division, with revenues surging 79% to 8.8 billion yuan. The tech giant's Kunlun Xiang P800 processors are making waves, delivering powerful ten-thousand-card clusters that recently trained the latest ERNIE model. Meanwhile, Baidu Intelligent Cloud continues to dominate the market, securing contracts worth five times more than its closest competitor.

A fig tree farmer's e-commerce venture turned sour when a buyer used AI to fake evidence for a refund. Despite the merchant's expertise proving the image impossible, the platform sided with the buyer in minutes. This case highlights growing concerns about AI-enabled fraud in online marketplaces and its devastating impact on small businesses. Authorities are now implementing AI verification tools to combat such scams, but for this farmer, the damage was already done - he closed his shop in frustration.