What’s The Plan for Plan A Legislation?
AI 2040: Plan A offers a clear set of policy recommendations that the United States government should try to implement. Given that we seem to be on track for a rapid path to superintelligence, it seems increasingly important to ensure that the governmental infrastructure is in place to prevent us from being annihilated.
What is currently being done to implement Plan A? Is there any legislation on the table that I can point my local political representatives towards and say “please support this”? If not, what’s the plan to make that happen, and what (if anything) can people with political influence do to support that plan?
Plan A is far more complex than Plan S, and has many more points of failure (even aside from the fact that Plan A envisions human handoff to AI in barely over a decade, an outcome I consider indescribably abhorrent). The momentum building the last few weeks points much more at Plan S than Plan A. Why not leverage that momentum to achieve Plan S instead?
Okay, sure, let’s go with Plan S. What are the concrete legislative proposals I can point politicians and high-profile political donors towards and say “please support this”?
In the US, the main bill to push for is the Sanders-Casar bill (the Ban Artificial Superintelligence Act).
Additionally, the bipartisan Lieu-Moran bill (the AI Kill Switch Act) is insufficient, but a useful step in the right direction.
In the UK, the Artificial Superintelligence Security Bill introduced by MP Alex Sobel.
The main case for Plan A vs Plan S was the idea that The Deal against all AI progress can’t last for millenia. Then the deal would dissolve and have the world re-choose whether AI progress moves through Plans A,B,C or D. And that’s ignoring the problems posed by the rise of China and disproportionately hardening a future race as opposed to the one where China merely smuggles chips or distills Western models.