There is a skeptical argument within epistemology that it is possible that you are merely a brain in a vat, somewhere in the area of Alpha Centauri, being fed all of the stimuli you think you are now having in the “real” world. In reality, you are living a life merely of brain and impulses, cleverly disguised in every way and in no way discernible from reality. You know nothing about the external world, the argument goes, because all of your knowledge has been fed to you and may be in no way connected with external reality (and if it were, you could not be said to truly know that). And you cannot prove that you are not now a brain in a vat, so you cannot prove that you know anything about the external world.
There are arguments against this, such as the epistemological dogmatism of Pryor (the thought that you can be immediately justified in believing things, like “I see a bright light”, based on your experience/perception of things, like seeing a bright light, and though these justifications might be defeasable, you are still immediately justified in them, and may come to knowledge based on them). Regardless of such arguments against this skeptical scenario, still the force of the argument remains: What if we were merely brains in vats? What difference might that make?
Video games, and particularly serious games (games intended to have the player learn something about external reality) are an interesting addition to this query. As 3D video game interaction grows in its pervasiveness and holistic mimicking of aspects of what reality is, players are further immersed into a reality that is constructed for their brains and for impulses from their brains. One might imagine a truly interactive serious game in which the brain is not mediated by the rest of the body, but is directly manipulated by impulses that allow the player to experience reality within in a wholly realistic way. Can knowledge gained in the brain in a vat skeptical scenario be transferred to knowledge in external reality? Most philosophers think not (whether or not they are correct is an open question).
James Paul Gee, author of a great book entitled What Video Games Have to Teach Us About Learning and Literacy, argues that video games (of the more modern variety anyways) are capable of teaching us many important things, not only about learning, but about how reality works. Embodied action can be transferable given reflection on problem solving. The question is “Is embodied action in games immediately transferable?”, i.e. do the knowledge, skills, and attitudes learned as a brain in a vat transfer to the real world? We might respond: “Yes, duh!” Stories could not teach us anything otherwise, case studies would be worthless, examples would not make sense, and pilots would not be trained on flight simulators if this were not the case. Perhaps the real question is “How closely must the brain in a vat environment match external reality in order for transfer to occur?”, i.e. useful knowledge, skills, and attitudes that are applicable in the world outside of the serious game. I would venture to say that it depends upon the risk involved in failing to transfer. If the risk is high, like the case of the pilot-in-training, who might plummet out of the sky due to insufficiently accurate in-simulation training, the learner should be put in situations as close to the real thing as physically possible for training (though without the risks involved in on-the-job training). The brain in the vat should have higher fidelity to the real world in cases of high risk. In cases of lower risk, this fidelity might be made lower.