Shanghai Artificial Intelligence Laboratory launches SafeClaw agent platform

On March 28, according to the Science and Technology Innovation Board Daily, at the Global Developer Pioneers Conference 2026, the Shanghai Artificial Intelligence Laboratory launched the SafeClaw agent platform, with "endogenous security, industry-level intelligence" as its core capabilities. In terms of endogenous security, SafeClaw builds a safe and trusted monitoring platform to achieve multi-level risk identification and interception; in terms of industrial-level intelligence, SafeClaw builds a multi-level agent super factory to achieve sub-second agent production based on hardware-level sandbox isolation, and performance is more than 10 times higher than mainstream architecture.