Playing With HAL

One of the banes of the modern MMORPG is the bot.

They pop up everywhere, but you see them most often where the rewards of a long grind are the greatest. If an extremely rare item drops off a type of NPC in a set location once in a million kills, there are likely to be bots in that area taking the burden of the million kills of the gold farmer on the other end.

For the most part these types of bots are pretty easy to detect. They tend to follow a fairly simple preset pattern, which even to a casual observer doesn’t appear like a human is driving. Humans, though somewhat predictable, tend to be fairly random in how they operate. Even a sophisticated bot will have patterns that appear over time to an observer, revealing their unhumanity. You called call them AI, but really only a very rudimentary form.

I was wondering then, if a MMORPG would be a good place to conduct a competition along the lines of the Loebner Prize. The Loebner Prize for Artificial Intelligence is a competition where entrants attempt to develop a program that can pass the Turing Test and respond to questioners in a way that they cannot tell if it is a human answering or the AI.

An equivalent test set in a MMORPG environment  would require that the contestant build an AI that would play the game in a manner indistinguishable from a human player. In fact this would be a tougher test than the Loebner prize since that contest only requires a conversation to appear human.

Interestingly, since most human players don’t farm gold in the intense way that the current generation of bots do, then if a contestant bot ended up gold farming in the same way, then it would be instantly recognisable as a bot and lose the contest.

On the other hand with the recent bot banning in Runescape, bot developers may want to create a more human bot to avoid such punishments in the future. They’d be less productive, but more secure. There’s a lot of money in gold farming, are these developments already underway?

It also seems to me that any serious AI R&D would be already be looking at the virtual worlds of MMORPGs to test out their experiments. It’s evolutionary after all, survival of the fittest, and MMORPGs can be pretty cut throat places. If an AI gets recognised as AI players report it and it is banned, killing it off, leaving the ‘fitter’ AI’s to fill its niche.

I wonder if this has already been done. Are we playing with HAL?


Please follow and like us:
Follow by Email