Discover the AI Tools

The AI community is buzzing after Brazil's promising Rio 3.5 model faced serious allegations just hours after its celebrated launch. Independent researchers claim the supposedly groundbreaking model actually combines code from two existing projects without proper attribution. The controversy raises important questions about innovation and ethics in the fast-moving world of AI development.

A Shanghai court recently sentenced a man who used AI software to create fake medical records and restaurant illness claims. Yang's scheme unraveled when restaurant owners compared notes about identical extortion attempts. Police easily spotted the forged documents, which contained impossible medical diagnoses and incorrect hospital stamps. The convicted scammer received four months detention (with probation) and a 2000 yuan fine - ironically matching his fraudulent demands.

Anthropic's powerful new AI model Claude Fable5 was abruptly withdrawn just days after launch when US officials declared its 'jailbreak' vulnerability a national security threat. The unprecedented ban affects both domestic and international users, raising questions about how we balance AI innovation with security. The company insists the risk was minor, but leaked reports suggest deeper concerns about potential foreign access to this cutting-edge technology.

Baidu's enterprise AI assistant DuMate just got a major upgrade that slashes operating costs by 75% without compromising performance. Through innovative engineering and its proprietary Harness engine, the Chinese tech giant has overcome one of the biggest hurdles in AI adoption - high token consumption. This breakthrough could make enterprise AI solutions significantly more accessible for businesses looking to deploy AI at scale.

Anthropic, the AI company that's been warning about its own technology's risks, now faces unexpected consequences. The U.S. government has slapped export controls on the firm after its CEO repeatedly sounded alarms about cybersecurity threats. While some experts call this an overreaction, others say the company brought it upon itself by stoking unnecessary fears about AI capabilities.

Tencent's AI assistant Yuanbao has joined forces with IMA's extensive public knowledge base, offering users direct access to specialized content across 20+ industries. Now when you ask Yuanbao a question, answers come with verified sources and one-click access to full articles. This partnership marks a significant leap in making professional knowledge more accessible through AI tools.

A Brazilian AI model developed by Rio de Janeiro's IT company is under fire for potentially being a 'frankenstein' of existing technologies. Technical analysis reveals Rio 3.5 may contain significant chunks of code from Nex N2Pro and Alibaba's Qwen3.5 without proper attribution. The discovery has ignited debates about ethics in the open-source AI community, with the Nex team responding with surprising restraint. This incident raises important questions about originality in an industry racing to develop powerful AI tools.

Ant Afu's upgraded 'Scan Skin' feature now identifies over 100 skin conditions with 99% coverage of common issues. The game-changer? Users can now get their AI analysis double-checked by real dermatologists from top hospitals - for free during the pilot phase. Early data shows remarkable 90% consistency between AI and doctor assessments, offering peace of mind alongside cutting-edge tech.

ByteDance's Douyin AI assistant just got smarter with its new Task Mode feature. This upgrade transforms the chatbot into an efficient digital assistant that can tackle complex projects. Need to research a topic and create a presentation? Task Mode breaks down your request, scours the web for relevant information, and even generates polished PPTs - all with minimal input. Industry watchers see this as a significant step toward making AI truly useful in workplace settings. While many chatbots entertain, Douyin is now aiming to become an indispensable productivity tool.

Science fiction author Hao Jingfang has ignited controversy by revealing half her new children's book series 'Galaxy Academy' was written with AI. While editors praised the work's quality, the admission has divided readers - some decry it as abandoning creativity, while others celebrate technological progress in storytelling. The award-winning writer, known for embracing innovation, stands at the center of a heated discussion about AI's role in literature.

Global banks are locked in a fierce competition for Chief AI Officers, with salaries reaching $3.5 million annually. An IBM survey reveals executive appointments surged 50% in just one year. But industry insiders question whether these high-profile positions will last as AI becomes mainstream in financial services.

WeChat Pay just supercharged its AI developer tools with version 2.0 of its AI Access Toolbox. The upgrade brings native support for nine languages, financial-grade security checks, and real-time troubleshooting - all designed to make payment integration smoother and safer for merchants worldwide. With these improvements, WeChat Pay is bridging the gap between AI assistants and real-world business needs.

Qwen AI Glasses have dominated the smart wearables market in early 2026, capturing over 30% of online sales. Their secret? A blend of cutting-edge tech like battery swapping and continuous software updates that keep users engaged. In a crowded field of 356 competitors, these glasses shine with features that actually solve real-world problems.

WeChat Pay has rolled out a major upgrade to its AI Access Toolkit, now supporting nine languages and slashing token consumption by 50%. The 2.0 version builds on merchant feedback to enhance security, efficiency, and usability. New features include financial-grade code review and real-time troubleshooting without leaving the coding environment. For global developers, this means smoother integration and significant resource savings.

Anthropic is rolling out mandatory identity verification for all users of its Claude AI platform starting July 2026. The process involves submitting government-issued ID along with real-time facial scanning through third-party service Persona. While raising privacy concerns, the company assures data won't be used for model training. This move reflects growing industry pressure to prevent AI misuse while balancing user convenience.

OpenAI is betting big on enterprise AI adoption with a $150 million investment to build a global partner network. The initiative teams up with consultants and tech firms to help businesses integrate AI into complex workflows. Early successes include payroll processor Paychex cutting wait times by 80% through AI-powered process redesign.

UB Tech's new U1 humanoid robots have taken pre-sales by storm, racking up 3,800 orders in just 10 days. Offering lifelike interaction through advanced motion simulation and a unique 'raising-style' emotional AI, these robots blur the line between tool and companion. While their adult-only policy raises eyebrows, the overwhelming response suggests we're entering a new era of human-robot relationships.

AI unicorn MiniMax faces intense scrutiny after abruptly changing its pricing model, leaving developers reeling from unexpected cost hikes. The company's shift to token-based billing for its flagship M3 model – without warning – has seen some users facing 257% price increases. This comes at a precarious time as MiniMax prepares for a major share release and intensifying competition from rivals like DeepSeek. The incident exposes deeper challenges in balancing R&D costs with customer expectations in the rapidly evolving AI market.

Apple is reportedly developing a powerful system-level AI assistant that could automate complex tasks across iPhones, iPads, and Macs. Unlike current AI tools with restrictive usage limits, this new feature might be bundled with Apple One subscriptions. However, significant privacy and security challenges remain before this ambitious project can become reality.

iFLYTEK has launched its groundbreaking X2-VL multimodal AI model, trained entirely on domestic computing infrastructure. This technological leap demonstrates China's growing capabilities in AI sovereignty, with the model scoring an impressive 148/150 on national college entrance math exams. The Wuxi-developed system combines innovative architecture with localized computing power, offering new possibilities for industrial applications.