unforum

[Silent]

Experiment Into Collective Intelligence

[Rogi Ocnorb]

*Rogi looks up the current addresses for Donald and Bettye Harden to ask if they'll be on his team.

[elfis]

Re: Experiment Into Collective Intelligence

[FLmutant]

Mind if I play along, Silent? It would seem to test that hypothesis you'd need to make sure:

1. There was a real solution possible (giving them the Voynich manuscript will always produce failure.)

2. Since it is collective, it needs to be something that isn't really solvable by any one person (so that it doesn't become a "can the community recruit the right resource" test.)

3. It can't be something that just requires group participation, or it isn't really testing the "intelligence" part (ie, SETI@Home isn't intelligent, it is just efficient.)

The more interesting question to me is WHY it fails, Silent. Not because I don't believe it exists, just because I believe it is an emergent property of something else. Maybe. It might just be self-organizing and not emergent. Or it might be emergent and not self-organizing.

I suspect it will fail because of the trust factor: the inherant belief that a fictional construct is created "solvable" produces a certain level of tenacity in pursuing that solution. I'm not sure how you would prove that without having a control -- one with a fictional construct that hid the fact a real CI task was hidden inside of it, and one without such a fictional construct where the CI task is explicit. I don't have direct evidence to support this idea except for the implosion rate in the games and what I generally describe as the "pirhana factor" (where the audience that were advocates become the fiercest critics.)

Interesting thought experiment in the very least: my gut says the emphemeral "CI" is more of a description of collaboration in general (any musician or performer can talk about "up time" where non-verbal communication between an entire ensemble and audience affects where the artistic output goes.) For many people, that experience of collaboration seems new ... or seems larger than it is because of media forms that spread out that experience over time.

Collaboration junkies

[vpisteve]

Hmm, from my perspective you're looking to disprove the existence of gravity, but hey, whatever floats your boat.

ETA: From reading your subject and subject description, you're really not clear on exactly what it is you're trying to disprove. The existence of a distributed hive mind that can very quickly solve complex problems, or an ARG that can solve 9/11??

If that's the case, the first one's gravity to me, while the second one is a perpetual motion machine.

[Rogi Ocnorb]

Would it also be good to make the task one that doesn't allow for ANY hints, helps, pushes or other interactive communication?
Strictly a task with a well-defined and measurable goal for success.
That would do away with any possible redefinition of the final goal due to various "leveling" activities that might occur during the run and reduce any emotional bias or hurt feelings in either direction.
Or... Is that a given?

[notgordian]

A study of this sort really does need clearer boundaries -- some limited forms have already occurred.

For instance, Perplex City has posed the "Billion 2 One" challenge (currently unsolved) and semi-jokingly asked for a proof to the "Reiman Hypothesis" (also unsolved).

On another note, Simon Singh released a series of 10 encrypted messages to go along with "The Code Book" -- although not collective intelligence per se, it involved a number of different teams tackling the same problem concurrently and resulted in improving algorithms.

The problem is, a sufficiently challenging puzzle must evoke enough interest to draw the attention of a "collective intelligence" -- and ideally it would need to be designed by a "collective intelligence", since the assumption is it should be do-able but beyond the capabilities of a single person, no matter how smart.

In fact, designing the problem itself would probably be the ideal assignment for a collective intelligence as proof of concept. Recursive, ain't it.

[krystyn]

Collective intelligence made me better at Halo.

[Silent]

Oh dear...just before I even send the trailhead and the experiment probraly going to collaspe.

FLmutant: I don't think I want to have a problem that "only" collective intelligence can complete (since if that is so, then the experiment would be in bias of CI, since CI is the only way). Since I am trying to disprove collective intelligence, I would rather prefer to have a very obscure problem that is based in "real life" situations, and see how the players solve it.

I also do think that a group of person, intelligent enough, can solve many problems, but I doubt that's really CI. I want to see if a random population can solve the problem. What I am worried is the possiblity of one intellignet person in that random group that will become the leader of the CI and automatically solving the real life problem, but the problem is hopefully obscure enough that it won't happen.

Another thing is the fact that in order to make this test, I cannot come up with a 'real' "real life problem" that has an answer (if I did, then it wouldn't really be a real life problem, would it ). In fact, techincally, the "real life problem" is a stand-in or an example of a real "real-life problem", in order to serve as an example. This "real life problem" does have an answer. I'm not going to try and have this ARG solve global warming, instead, I'm going to have players deal with a complex problem with an answer, and see if they get it or not.

vpisteve, I am looking to disprove the existence of an effective distributed hive mind that can very quickly solve complex problems correctly. I'm betting that a hive mind won't solve complex problems, but that's why this experiment is being done.

Rogi Ocnorb, it was expected that once the player gets the "real life" problem, that's it. It's up to them to solve it. If they solve it correctly, congrats. If they don't, then they suffer the consquences. In this ARG, failure will not be tolerated.

notgordian, recursive loops are exactly the reason why I wish to disprove this theory.

[vpisteve]

Well, then your main hurdle is going to be able to build up an adequate player-base made up of motivated, engaged participants to provide a sample large enough to validate whatever result you get.

[catherwood]

[mapmaker]

[FLmutant]

[Rekidk]

We need to get Jane McG in here; isn't she, like, the queen of Collective Intelligence?

[imbri]

I'm gonna jump on the backs of the others here and say that if you want to "prove" anything (even to yourself), you're going to need to make a more sound hypothesis and that, as FLmutant points out, includes defining the term Collective Intelligence as you see it. It sounds to me like you want to focus on the problem solving aspect of intelligence but that leaves out things like the ability to acquire & apply knowledge.

In ARGs, I think we do see collective intelligence at work. I know that we see collaborative intelligence, but I do think that there's some collective intelligence in there as well. And, while you state that you are worried that one person with the right knowledge would come in and solve your puzzle, that is a part of collective intelligence. We all have our own skills, knowledge, and strengths and when we come together as a collective we are able to better harness those things - and that's been another thing that has always impressed me. I still get giddy when I think of the way Steve and DonkeyOatey sought out help solving a cypher and wound up bringing one of the best cryptographers around into the community. Though, I suppose, that's more an example of collaborative intelligence.

That said, I don't want to jump on the collective intelligence bandwagon. I do have my doubts, but I must admit to a certain excitement about it all. The hippie freak is very much attracted to the stuff Tom Atlee's been doing & thinking and I would recommend his Tao of Democracy to anyone with an interest in political change. Though, I suppose, Pierre Levy's Collective Intelligence is probably of more interest to most people here. Or, of course, Surowiecki's Wisdom of the Crowds which is more pop than philosophy (and filled with problem solving sorts of examples)