Formulate AI-super-projects that would be both prestigious and socially beneficial
Artificial Intelligence, Great Power Relations
There are already some signs of race dynamics between the US and China in developing TAI. Arguably, they are at least partly motivated by concerns of national prestige. If race dynamics speed up, it might be beneficial to present a set of prestigious AI-projects that the US and other countries can adopt. These projects should have the following features:
Be highly visible and impressive for a wide audience
Contribute to safer AI (e.g. through interpretability or alignment playing a great role in the project)
Be socially beneficial (i.e. the benefits should be distributed widely, ideally the technology would be open access after development)
Formulate AI-super-projects that would be both prestigious and socially beneficial
Artificial Intelligence, Great Power Relations
There are already some signs of race dynamics between the US and China in developing TAI. Arguably, they are at least partly motivated by concerns of national prestige. If race dynamics speed up, it might be beneficial to present a set of prestigious AI-projects that the US and other countries can adopt. These projects should have the following features:
Be highly visible and impressive for a wide audience
Contribute to safer AI (e.g. through interpretability or alignment playing a great role in the project)
Be socially beneficial (i.e. the benefits should be distributed widely, ideally the technology would be open access after development)
(Idea adopted from Joslyn Barnhart)
Possible downside: Contribute to further speed-up of AI development, possibly leaving less time for alignment research
(However, if done correctly, this project only harvests pre-existing dynamics and leads funds to beneficial projects.)