Anthropic has released a report highlighting cybersecurity concerns over the AI model GLM-5.3, developed by Chinese company Z.ai and made publicly available since August. According to MIT Technology Review, the model's open download and modification capabilities could be exploited in cyberattacks.
GLM-5.3’s free accessibility contrasts with industry trends toward more controlled AI development. Nvidia CEO Jensen Huang recently advocated for the United States to increase efforts in creating more open AI models, emphasizing the strategic importance of AI innovation, as reported by MIT Technology Review.
For Japanese investors and market participants, the release of openly accessible AI models like GLM-5.3 underscores the growing global cybersecurity risks tied to AI technology, highlighting the need for vigilance in tech-related equities and regulatory developments.