Because mostly it is a game engine at its core. The rendering pipeline is one the core elements which is hard to add features with low probability of broken things. That would required year men hours of development. Anyway some progress are being made with RTX which goes towards movie making and Im quite sure will succeed for visualization aswel. Don’t be fooled that these features are here because they will cook as experimental for quite long before they reach the production statuses where you can grab this or that functionality and apply to gaming. When it is a common ground for graphics cards to process at 60+FPS with a power of 30TFLOPS with a price suitable for gamers, any improvement on GI would only benefit film and visualization with extremely expensive hardware.