I am afraid that I will have to fall on the side of go here. Pardon me if I am wrong, but the essential differences between the two games are very much derived from the essential differences between Western philosophy (specifically Islam, Judaism and Christianity) and Eastern philosophy (specifically Buddhism and Taoism). Victory is accomplished for the most part not by a flat-out rout or a sweep of doom across the board, but by slow and careful play. In addition, the concepts of (for example) potential territory, life and death, and sente and gote (that is, the advantage) have a way of emerging from the position around them. They are not (for the most part) explicit in the stones' positions, and yet they emerge from them, creating a second metaphorical board on which they are expressed. And emergent properties, of course, are my lifeblood. Moving deeper into the emergent board, the pseudoboard that is created from the real one, we find other concepts -- for example, aji ("taste" left behind in an unfinished sequence), miai (interchangeable points), and influence. Of course, chess illustrates a few emergent properties as well, most notably open files and influence in the center. However, the pieces can move in chess, and already one layer of possible epiphenomena, or emergent properties, is destroyed. In go, the motion is never actively forced, and yet it happens: walls of stones may race to the edge, helpless near-prisoners may jump for defense.
Another argument in favor of go is the versatility of the equipment. Chess pieces and boards, though often beautiful and artistic, can effectively only be used for one game: chess. Even checkers has more versatility and possibilities. The go board, although simple, has been used for many games in recent years, notably some created by John Conway. The simplistic board and stones have been used to create, without any other equipment, the games of Gess, Phutball, and Epaminondas, to name a few. With a different board (probably a go board) and some creative interpretation of the pieces, chess can give birth, perhaps, to Chinese chess.
Finally, there is one major difference that has recently separated chess from go. This is, of course, the birth and death of computer chess. By the use of very large memory reserves and a dynamic variable crudely calculating the value of each position, the computer Deep Blue was able to defeat Garry Kasparov in competition. Let us call this kind of computer a number-cruncher: the key to playing a game is being able to calculate very deep and fast. This number-cruncher, in effect, managed to knock computer chess away from the field of study.
Go, meanwhile, is still standing strong as a field, even being used as a testing ground for young AI hacks. Maybe this has to do with the size of the board: after all, the go board is four times as big as the chess board, and attack is much harder to accomplish than in chess. However, the versatility of the pieces means that computer go can be used to try out new techniques: for example, another type of program, a learning program, can be used to evaluate joseki. Yes, indeed, computers are slowly learning to evaluate high-level epiphenomena, to make another connection. The majority of programs are probably 5 kyu level, and yet they can evaluate the influence of a group. This, in my opinion, is more important than beating a grandmaster, because it means that the dawn of computer go should come hand in hand with the dawn of artificial intelligence, and not just with bigger, badder hardware. Because of all these factors, go has the ability to be a real metaphor for cognition, which is just what the new generation of AI programmers needs.