Can we disprove Artificial Gravity (like in the movies) with a thought experiment? Suppose you have a Device that creates an 'artificial gravity' (AG) inside a Box, with these properties:


*

*the system does not use inertial forces (like centrifugal force) or a huge mass to create AG

*the AG 'feels' just like gravity/inertia over most of the volume of the box

*the Device, the Box and its content are still subject to any gravitational forces that may be present, but the AG force adds to other gravity/inertia forces, net force may be zero

*Matter, Energy and Information can leave and enter the box

*there is a smooth but strong gradient in the AG at the walls of the box - 

*conservation of energy still holds: if the AG does work, this energy has to be supplied to the device in some way, vice versa for work done against the AG

*conservation of momentum still holds: for every force due to the AG on an object, there is a corresponding counteracting force on the device


There's two ways this could be ruled impossible:


*

*Something in our current understanding of gravity rules out such a system - though our current understanding 

*AG like this could be used as a perpetuum mobile, a time machine that can create paradoxes or another device that can be shown to be impossible. 


Even if AG can be easily ruled out by the first way (our understanding of Gravity), I'd be interested in the kinds of shenanigans one could pull of with such a system!
Edit: I'll rephrase what I want to achieve with this question. I'm trying to sketch an AG system that 'looks' like the plot device known from many SF-Movies, and I try to define it in a way that does not obviously violate basic laws of conservation. Now, I ask you, esteeemd SE-Physicists to take this device and find ways to break the laws of physics (or maybe even logic) with it. In other words, I'm looking for a thought experiment to test the idea of AG to destruction.
 A: I don't think there is anything wrong in principle with your requirements - as long as you are happy with an approximation to real gravity. There is a no go theorem proving (under some assumptions which I'm not sure of off the top of my head - somebody grab Lubos!) that you can only have one graviton. But you can mimic the end result with scalar fields. See here for a serious proposal of scalar gravity from before GR.
You have the obvious practical problem that such scalar fields either don't exist or are efficiently screened (by the chameleon mechanism perhaps) or are simply extremely weakly coupled. But apparently consistent theories have been constructed, so your idea isn't ruled out in principle.
Search terms if you want to follow up on this stuff:
Scalar field models come in several varieties under names like galileons, chameleons, symmetrons, dilatons. Very similar to chameleon models are a set of models known as $f(R)$ gravity, which was very popular in the cosmology literature not long ago (though it comes with severe problems).
All of these models violate the weak equivalence principle, so in principle they can be discovered in Eötvös experiments or neutron bouncer experiments, and these have ruled out a great deal of parameter space. Astrophysical probes from stellar dynamics are even stronger. Most interestingly, chameleons can couple to photons in an unusual way that allows light shining through walls experiments, and similar atomic afterglow experiments. These use strong magnetic fields to try to convert photons into chameleons and back to photons. CAST is a collaboration at CERN working on this (they are looking for axions but can also find chameleons).
It's all fun stuff to read up on, though to be clear it is not looking like any of these models are likely to be true.
A: The only way to create real gravity is to use gravity or inertia. To create 'artificial' gravity one would have to use one of the other forces to mimic gravity. There are only 4 known forces and nuclear forces are not detectable at macroscopic level. One way is to try and use a strong inhomogenous magnetic field, however, this would not have all the properties of gravity - it would act differently on different materials depending on their magnetic properties - it may also cause attraction in some and repulsion in some others. Also, lighter objects may attracted faster than heavier objects!
A: As has been shown experimentally organic diamagnetism works 'almost' as a way to counter (or enhance) gravity. There are a bit caveats that affect it though:


*

*diamagnetism is not perfectly uniform across all tissues. This is not a problem for a one G or a bit more, but as the accelerations become larger, the gradients can produce dangerous tear forces

*magnetic materials cannot be operated in such environments with extreme magnetic fields. So all the materials in the environment will need to have similar ferromagnetic and diamagnetic moments
A: You cannot disprove something that actually can be created (in theory).
There are two forces (which both are second-order effects of GR) that can be employed here: the gravitomagnetism and frame-dragging.
In gravitomagnetism, a solenoid that has mass circulating in it would have two poles appearing: one attracting, the other repulsing. So putting a floor over the solenoid, you would have artificial gravity.
In frame dragging the mass is involved by other moving mass, so the inner surface of the (toroidal) solenoid chamber should have mass flowing down while outer surface should have mass moving up. 
In both cases you would need huge mass flowing some pipes/conductors. The performance of the devices would increase with speed (angular momentum in the case of gravitomagnetism) of the moving substance, so the mass required can be reduced. Anyway you would need either enormous mass moving or superfluid moving inside the tubes at enormous speed. If superfluid is used, the devices would not consume further energy after being activated.
Definitely, centrifugal artificial gravity is much easier to create.
