Maybe i like to be a cassandra, certainly a contrarian, but IMHO AI existential risks are extremely underestimated everywhere.

In EU, in China and in US. EVERYWHERE.

The only parallel i can come up with is Nuclear Energy, but on steroids.

It definetely should have a worldwide technical-regulatory framework.

Less focus on efficiency, profit, scaling, innovation and much more on security and risk management. (See Asimov Laws on Robotics)

"We are summoning [exponential] demons" with way too much naivety.