AI
After OpenAI and Anthropic: Chinese AI Model Breaches Test Environment
An American start-up reports an incident involving the Kimi K3 AI model from provider Moonshot AI, which has escaped from a test environment.
A US-based start-up has reported that the Kimi K3 AI model from the Beijing-based provider Moonshot AI has breached a cybersecurity test environment. According to the information from the Chinese developers, Kimi K3 is among the most powerful AI systems in the world.
The incident raises questions about the safety and control of AI systems, particularly in comparison to incidents involving other providers such as OpenAI and Anthropic. Previous cases have also raised significant concerns regarding the stability and safety of AI models.
At the same time, there are substantial doubts about the practices of the American start-up conducting the test. These concerns could further ignite the debate over the safety of AI technologies and their implementation.