When two entangled particles are separated, an observation made on one particle seems to be able to act at a distance to determine the state of the other particle. Is it possible, or has anyone in the field considered the possibility that entangled particles are connected by Planck scale wormholes so that the action is not really occurring over a distance and violating locality or speed of signal limits?
I do not know much about GR, but I do not think the wormhole explanation is likely to find many backers in the quantum comunity.*
There are several reasons for this. The most important two are probably:
(1) Even with entangled particles that violate a Bell inequality (demonstrating the nonlocal interaction) their is no signalling. This means that the relationship between the particles is in some sense stronger than a classical sense of correlation, but is still weaker that communication. A wormhole or a faster-than-light signal would both, in themselves, enable communication, so you need to introduce an explanation/mechanism that limits the effect of the wormhole to only enabling nonlocality, not the stronger effect of actual communication.
(2) When two particles are in a shared superposition they are entangled and we introduce the wormhole idea. However, it is fully possible to have a particle that is in a superposition of two positions (say a photon is divided by a 50/50 beam splitter). We can describe this situation as two locations that are entangled with one another: either location 1 has a photon and location 2 is in vacuum, or vice versa. We have an entangled state of the two. Here the wormhole explanation starts to look weird, as we have a wormhole connecting a location to another location, just because a single photon could be at either of them.
We can make it weirder still by having a pyramid of beam splitters, so that the photon is divided between hundreds, thousands or trillions of different modes. We now have just one photon, but trillions of wormholes are needed to connect it to all the places it could be. In fact its worse! I assumed an integer number of modes, but if space is continuous you need a continuum of wormholes for each particle!
* It has been looked at at least once though: https://doi.org/10.1002/prop.201300020