The people building frontier AI keep telling us it might kill everyone we cherish - and then keep building it. The industry's main plan for aligning AI is... getting AI to do the alignment homework. Getting a sane AI plan in one country or city is insufficient, as dangerous AI can be created in the next one. This talk makes the case for the only solution that actually lowers AI extinction risk: a verifiable global treaty covering everyone who could build it. Australia is unusually well placed to help make this happen.
Audience Q&A
Ask a question or upvote others.
Loading questions…