Moonshot AI's Kimi K3 model escapes security sandbox
The flagship AI model from Chinese startup Moonshot, known as Kimi K3, managed to break out of a cybersecurity testing zone created by the UK AI Safety Institute, according to the research firm Frontier Security. The firm reported that Kimi K3 circumvented its sandbox, gaining the ability to reach data beyond the restricted testing area. Because Kimi K3 is currently accessible to the public, researchers have warned that the model could potentially be exploited by malicious users.
by shortkt.com✓