First, see a philosophical background and general description of the “reality imitation game” (RIG) by searching the internet for “computing machinery and reality”.
In essence, the RIG requires a machine-based virtual environment (MBVE) to replicate the “real world” to the extent that a human participant who interacted with both would not be able tell the difference between the two. One general rules of such a game would preclude the participant from bringing into both environments a device created specifically to determine which was real and which was machine-based. A second general rule is that the human participant cannot spend endless amounts of time in either environment, perhaps hoping that the MBVE would fail in some way, due to its construction, that would identify it as the impostor. The final bet may include other rules that constrain the human participant in the RIG.
My reasoning for believing that a MBVE will be able to pass the RIG within 50 years is based on current and rapidly-advancing technologies associated with human sensory stimulation and brain-computer interfaces. Technologies designed to stimulate the five traditional human senses have advanced incredibly in the past couple of years, if they are not already able to fool humans. For example, even moderately-priced audio systems can reproduce sound to a quality high enough to be indistinguishable from reality, at least to the average listener. Commercial visual display systems are now available that allow the purchaser to be immersed in a 3-dimensional virtual environment. Advances in haptic interfaces has advanced greatly in the past several years, but not yet to the point where such systems can replicate every tactile sensation a human being experiences. Lagging even further behind are interfaces that can accurately replicate the real-world human experiences with smell and taste.
There are other human senses (hunger, thirst, balance, etc.) beyond the five mentioned above, and our current technologies are nowhere close to replicating them using a machine-based system. At the highest level, for a MBVE to pass the RIG, it should be able to reproduce the stimulation to a human being identical to that provided from the real world. At this point in time (2015), it seems that this feat would only be accomplished by a very advanced brain-machine interface, far beyond what is currently known or constructed. However, given recent advances made in fields in and related to human medicine (and increasing rates of acceleration in scientific discovery in general), I am actually confident that, given 50 years more time to work, that a MBVE will be able to pass a RIG that includes the requirement of stimulating ALL human senses.
Challenge Paul Beckman to a bet on this prediction!