In a statement released on September 16, leading AI scientists warned that AI could cause serious harm if human control is lost. The scientists said countries need to develop authorities that can detect and respond to AI incidents and that a global contingency plan should be established.

In the long term, states need to develop an international governance regime to prevent the development of models that pose global catastrophic risks. These views are based on findings from the International AI Security Dialogue held in Venice in early September.

We welcome your comments!