Editor's Blog
In 2009, we reviewed the results of the 2008 2K BotPrize Competition and raised the question “Was That a Bot or a Human?” Now the results are in for the 2010 competition: the ability for judges to distinguish between gamebots – those wily non-human AI online game characters – and humans – has narrowed almost to the point of standard error, a mere 3.6657% between the “most human” bot and the “least human” human. Here are the results.
Gamebots (as opposed to Internet bots or web robots) are a type of weak AI expert system software used to simulate human behavior in computer games such as Unreal Tournament and its ilk: World of Warcraft, Guild Wars, Lineage, and Everquest. The BotPrize Competition uses a specially hacked version of of the first person shooter Unreal Tournament 2004 so an AI program on a user’s PC can send sensory information for a character over a network connection.
This year’s most human bot – designed by the Conscious-Robots team of Jorge Muñoz and Raúl Arrabales – came very close to passing the videogame equivalent of the Turing Test. The Turing Test includes an interrogator or judge (Player C) tasked with determining which of two players (Players A and B) is a computer program and which is a human. The judge is typically limited to using responses to written questions in order to make the determination. In the case of the BotPrize, the judges actually played against the other players and then rated them.
The winning bot developed by the Conscious-Robots team (CERA-CRANIUM Bot 2 or CCBot2 for short) runs the CERA-CRANIUM cognitive architecture, which is based on the CERA-CRANIUM computational model of machine consciousness derived from global workspace theory. The global workspace model of consciousness, proposed by Bernard Baars, an Affiliated Research Fellow of The Neurosciences Institute in San Diego, California, proposes that perceptions below the threshold of consciousness are processed in relatively small, local areas of the brain. Broadcasting this pre-conscious information to the global workspace — a network of neural regions — results in conscious experience.
As Arrabales points out in a paper on conscious-like behavior in computer game characters, physical embodiment and real world situatedness are likely key factors in the production of “machine consciousness.” He argues that software agents such as video game bots can be “embodied and situated” with a digital body equipped with software sensors and actuators that allow them to perform actions.
What humans have that’s missing in bots is “a combination of cognitive capabilities like attention, learning, depiction, set shifting, Theory of Mind (ToM, or the ability to attribute mental states to others), planning, feelings (in the sense of higher order representations of emotions), etc.”
So, if a bot fools the judges in next year’s 2K BotPrize Competition, does it pass the Turing Test? Perhaps. Is it conscious? You be the judge.
Sign up for the Humanity+ newsletter:
3 Comments
UT is not really related at all to the MMOs you listed; it’s a deathmatch-style first person shooter, more akin to Half Life, Halo, Doom, Quake, etc. Bots in MMOs have to follow vastly different behavior patterns and are generally easier to spot because of the social interactivity and cooperative group play aspect of an MMO. Not to disparage fps bots or anything, giving them believable human characteristics presents its own kind of challenges.
I agree with Ilk. Though MMO’s are trying to get bot’s that are indistinguishable from human counter part’s. It’s still in the infancy stage. FPS is one type MMO’s are another type. FPS is less communication and more of a aim, move, taunt, fire, strategy. And MMO’s are more social like pet/companion/love interest, merchant/seller or buyer, story telling, teaching/guiding/leading, and a few others. Though both are based off human like trait’s, and are just a small piece of the bigger puzzle for the different types of sentience or human equivalent or greater A.I.’s. But that’s lot’s of pieces and many combinations to go threw yet. Medical psychological A.I. is one of the main lagging A.I. field’s right now for a puzzle piece like approach to good A.I. . DARPA has some of the other pieces being developed threw prize like event’s and collage funding. The rest are already done and exist ( Deep blue, stock market bot’s, medical A.I.’s, search engines algorithms, forecasting, chat bot’s, and a ton of others) Most we take for granted due to the simplicity. Though some thing that is reflected in human behavior and actions today. How they best come together and are streamlined for good A.I. is the race of today. Epic and Id’s FPS A.I. is good and soon will be indistinguishable from humans if not superior ( I.E. nightmare bot’s). But I would also stress this is one of the harder and slower way’s to get to sentient A.I. systems. Evolutional or emergent might actually be easier under the right conditions. Reverse engineering the brain is probably the hardest way, but with the tool’s and knowledge is still feasible today. Which ever way comes first it will not happen over night, most of this takes years of work and debugging or refinement with testing for a good level. Many way’s and many path’s to get to the end result though. So help assemble and rule in or out the different part’s, path’s, approaches, choices, to start to get us the end result’s that will be good and beneficial. I see a lot of pieces, but some are still missing or lacking. Just my 2 cent’s of relative perspective on this topic, and good to see some people step up as so much of the game industry has been stepping backward and dumbing thing’s down for consoles. Part of why I’m moving back to modding and buy so few games today, though I’m still looking for the right game engine for the high res. I so love.
I’ll be impressed when the bots can hold their own with smack talk on XBox Live…