It is crucial to consider how to ensure positive outcomes when building powerful AI systems such as AGI, similarly to the approach taken after nuclear weapons were created. Additionally, it is easier to support the negative side of AI than the positive side.