Hey, interesting topic. Strangely this was the subject of my (unfinished) masters thesis from many years ago. Just in the last few months I've come back to the idea and have been working on proving the concept in a dive light application.
Bottom line, using a fluid filled body will eliminate the possibility of water intrusion due to pressure effects. It's just basic physics. Bell & Ross (a high end watch maker) uses this in a watch that is pressure tested to something like 10,000 FSW. A Japanese submersible manufacturer that does super deep exploration uses liquid filled batteries for the same purpose...pressure resistance. Some fiber optic manufacturers also used oil filled lines to prevent deformation on ocean floor cables.
Since flooding is the #1 source of dive light failure, eliminating that possibility seems like a good idea, to me at least
Any light with proper seals should hold the fluid in without any issues...just like proper seals keep it out. Leakage would be the result of manufacturing, design, or user error. However, say you get some debris on the o ring and go for a dive. Also assume the debris would cause a flood in an air filled light. If it was filled with fluid, it still wouldn't flood because there is no place for the water to go. Pretty cool if you ask me.
As far as fluids, mineral oil is a common dielectric used for insulating and cooling high voltage transformers and switch gear. Silicone oil is another possibility. There are also a host of dielectric fluids used in liquid PC cooling. Finally, there is a proprietary fluid called Midel that is designed to be a stable dielectric insulator.
The biggest challenge in fluid filled dive lights is the design, as opposed to the physics. Hopefully I'll be working some of this stuff out over the next couple of months. I'm going to be using PC cooling fluid to start testing with but I don't think it will be chemically stable enough over the long term. My first test will be to determine the pressure generated by heating the fluid in a sealed environment and use this to determine how important additional pressure compensation is. I'll keep ya posted!