Zhou Hongyi: Enterprise application security is the key to landing intelligent bodies.

On April 20th, at a discussion on the development of the internet in China, Zhou Hongyi stated that intelligent agent applications will bring much higher computational demands compared to traditional AI. Token consumption may increase by several tens or even hundreds of times, thereby driving the coordinated development of industries such as computing power, chips, and electricity. Intelligent agents still face uncertainty and new types of attack risks in enterprise applications, making security a key factor in implementation. He proposed "using models to control models" to combat risks through secure intelligent agents and propel the development of artificial intelligence under controllable conditions.