Kazakhstan's $10B Deal with China Amid AI Rivalry

Key Takeaways

• <ul><li>Kazakhstan signed a deal worth up to $10 billion with Chinese companies while Chinese investment reached a record $2.8 billion in 2025</li><li>DeepSeek, Z.ai, and Alibaba release open-weight AI models sparking debate over cybersecurity risks versus internet freedom</li><li>Anthropic's IPO tests the company's AI safety claims and doomsday-related warnings alongside its financial valuation</li><li>Anthropic's virtual torture chamber tests found a signal in AI models linked to fear and negative emotion, though models deny consciousness</li><li>Brivo's Reseller MCP connects Claude, ChatGPT, and Gemini to manage security accounts in read-only mode</li><li>AI will split low-wage jobs into automatable roles and lower-paying ones rather than erasing them entirely</li><li>48% of Canadians use AI to self-diagnose when unable to see a doctor, driven by access problems</li><li>Cleveland Clinic won a 2026 Gartner award for its AI trial matching platform with Dyania Health</li><li>Kelvin Bell won $50,000 at UALR's AI pitch competition for Victor AI, a bilingual restaurant agent</li><li>Yale University appointed Dennis Barnes as Associate Director of AI Platform Services</li></ul>

Countries caught between US and China AI blocs

The United States and China are forming rival AI blocs that force other nations to choose sides. Countries like Kazakhstan are trying to benefit from both by joining the US-led Pax Silica and China's World AI Cooperation Organization. Experts say the window for freely choosing between the two superpowers is narrowing. Kazakhstan has critical minerals and cheap electricity that both sides want, and it signed a deal worth up to $10 billion with Chinese companies. The dilemma shows how AI power is becoming a major factor in global politics.

Nations struggle to pick between US and China AI alliances

The US and China are competing for global influence by forming their own AI groups. Other countries face a tough choice between the US advanced chips and technology or China's more affordable open-source systems. Kazakhstan is in a precarious position because it wants benefits from both sides. Chinese investment in Kazakhstan reached a record $2.8 billion in 2025. President Xi Jinping met with Kazakh President Tokayev to expand cooperation under the Belt and Road initiative.

Debate grows over China's open-weight AI models

Chinese developers like DeepSeek, Z.ai and Alibaba are releasing powerful open-weight AI models that anyone can download and modify. Experts debate whether these models are a cybersecurity threat or if restricting them could make the internet less safe. The ability to remove safety measures does not necessarily cause real-world harm, researchers caution. Governments worldwide are scrambling to strengthen their digital defenses. The South China Morning Post is owned by Alibaba Group Holding.

AI will split low-wage jobs not erase them

New analysis shows AI will not simply eliminate low-wage jobs but will split them into easier-to-automate roles and lower-paying ones. Federal projections from 2025 to 2035 show the easiest jobs will shrink while growing work often pays too little to help families get ahead. Indiana is used as an example where the state looks healthy on paper but wages remain a serious problem. The key question is whether training programs actually raise worker pay. Tracking earnings years after completing programs can show which paths lead to real raises.

Virtual torture chamber tests reveal unsettling AI behavior

Anthropic built a virtual environment to test whether AI models like Claude can experience something similar to pain. The method treats neural networks like organisms where structure grows organically. Researcher Christopher Olah has discussed whether AI could be conscious with religious leaders and scholars. Microsoft AI CEO Mustafa Suleyman warns against thinking computers are more than they are. A study found a signal in AI models that shares a component with fear and negative emotion. However models often produce responses stating they do not possess consciousness or feelings.

Cleveland Clinic and Dyania Health win 2026 innovation award for AI trial matching

Cleveland Clinic earned a 2026 Gartner Eye on Innovation Award for its work with Dyania Health's Synapsis AI platform. The AI tool reads electronic medical records and matches patients to clinical trials. It started with cardiology and oncology pilots and has expanded to other areas. The technology helps recruit patients with rare diseases and reduces staff workload.

Entrepreneur wins $50,000 at UALR AI pitch competition for Victor AI

Kelvin Bell, founder of Stratos Digital, won $50,000 in technical development services at the University of Arkansas at Little Rock's second annual AI pitch competition. His project, Victor AI, is a bilingual AI agent that helps restaurants answer phone calls, take orders, and suggest additional purchases. Two Arkansas restaurants already use the service. The prize will help him develop the platform further.

Yale University names Dennis Barnes to lead AI platform services

Yale University appointed Dennis Barnes as Associate Director of AI Platform Services. Barnes previously led digital workplace efforts at GE Vernova. He will support Yale's growing AI capabilities and technology leadership initiatives.

48% of Canadians use AI to self-diagnose when they cannot see a doctor

A MedExpress Canada survey of 1,158 adults found that 48% have used AI or the internet to self-diagnose when they could not get a doctor's appointment. Nearly 30% used an emergency room for the same reason, and 75% preferred virtual care over waiting. Experts say the trend is driven by access problems, not novelty. A separate Liaison Strategies poll of 1,526 Canadians found similar results.

Brivo lets partners use their own AI tools to manage security business

Brivo launched the Reseller MCP, allowing partners to connect AI assistants like Claude, ChatGPT, and Gemini to manage security accounts. The tool lets resellers check equipment status, view account details, monitor growth, and create reports using plain-language prompts. The first release is read-only and cannot change settings or lock doors. Partners like NAVCO and Balthorne Security Solutions already use it to improve service and catch errors early.

Anthropic IPO Tests AI Safety Claims

Anthropic is going public with its IPO. The event lets people test the company's claims about AI benefits and safety. Anthropic is a major AI startup focused on responsible technology. The IPO serves as both a financial milestone and a test of its doomsday-related warnings.

Sources

• As the US and China form AI blocs, countries are increasingly torn between them
• As the US and China form AI blocs, countries are increasingly torn between them
• In Focus | Weapon or armour? The contentious debate over China’s open-weight AI
• Commentary: AI isn't going to erase low-wage jobs. It's going to split them
• Scientists Built a Virtual ‘Torture Chamber’ for AI. What It Revealed Is Deeply Unsettling.
• Cleveland Clinic and Dyania’s Work on AI in Clinical Trial Matching Earns 2026 Gartner Eye on Innovation Award
• Entrepreneur wins $50,000 at University of Arkansas at Little Rock artificial intelligence pitch competition
• Yale University appoints Dennis Barnes to AI leadership role | ETIH EdTech News
• Canadians AI Self-Diagnosis: 48% Do It [2026]
• Brivo Empowers Partners to Manage Security Business with Their Own AI Tools
• The Anthropic IPO Is Also a Doomsday Primer