Short paper

Is unregulated AI research a good idea?

The familiar pause-or-accelerate argument changes depending on whether AI development is an arms race, a prisoner's dilemma, or a tragedy of the commons.

PDF · 4 pages/AI governance/game theory/acceleration

I wrote this because “pause or accelerate?” is not one argument. It is at least three games with different losers: an arms race, a prisoner's dilemma, and a tragedy of the commons.

The paper is short and more interested in identifying which game we are actually in than in producing another confident sign for either camp.

If the inline reader does not load, open the PDF directly.

Discussion

Corrections, disagreements, and extensions are welcome.