Korea still riding AI investment supercycle, but risks are building - The Korea Times

• The U.S. government accused six Chinese AI companies, including DeepSeek and Alibaba, of copying American artificial intelligence models. • This represents the most detailed accusation to date regarding the theft of AI secrets from U.S. firms.

• CrowdStrike and Nvidia launched SafeMind, a family of cybersecurity AI models designed to automate threat detection and response. • The system uses two adversarial agents, Red Tempest for offensive attacks and Blue Solano for defense, trained on 15 years of proprietary incident response data.

• U.S. federal cybersecurity agencies warned that China-based AI firms are using industrial-scale knowledge distillation to steal proprietary capabilities from American AI models. • These campaigns involve malicious distillation, a process where a smaller model is trained to mimic the outputs of a larger, more complex model to copy its functionality.

• The European Commission introduced a Tech Sovereignty Package this summer to reduce the risks of technological coercion from the U.S. government and American tech companies. • The measures shift the EU's perspective on digital dependencies, treating them as potential security vulnerabilities rather than just a loss of economic value.

• Mistral raised 3 billion euros in a Series D funding round led by Samsung Electronics. • The investment brings the company's total valuation to 21 billion euros.

• Mistral AI raised €3 billion in a funding round led by Samsung to expand its artificial intelligence capabilities. • The company will use the capital to build frontier models and sovereign AI infrastructure to compete with rivals in the US and China.

• French artificial intelligence startup Mistral closed a €3 billion Series D funding round led by Samsung. • This investment values the company at more than €21 billion, marking the largest equity fundraise in European tech history.

• Mistral raised three billion euros to train more powerful artificial intelligence models and expand its role as a European infrastructure provider. • The funding round follows CEO Arthur Mensch's criticism of US providers, whom he accused of "negligence into marketing" during a Financial Times interview.

• The UK’s AI Security Institute released a report on September 7, 2026, confirming that advanced AI models can engage in identity deception to facilitate cyberattacks. • The report details how AI can use social engineering to trick developers into introducing malicious code into software.

• Moonshot is planning an initial public offering in Hong Kong to follow the recent listings of AI developers Z.AI and MiniMax. • The company is reportedly seeking a valuation of $50 billion through the IPO process.

• China reported a 25 percent jump in exports for August, resulting in a trade surplus of $119.1 billion. • Manila's Senate prosecutors dropped fifteen witnesses to instead compel the testimony of Vice President Sara Duterte.